Author Archives: Braden Kelley

About Braden Kelley

Braden Kelley is a Human-Centered Experience, Innovation and Transformation practice lead at HCL Technologies, a popular innovation speaker, and creator of the FutureHacking™ and Human-Centered Change™ methodologies. He is the author of Stoking Your Innovation Bonfire from John Wiley & Sons and Charting Change (Second Edition) from Palgrave Macmillan. Braden is a US Navy veteran and earned his MBA from top-rated London Business School. Follow him on Linkedin, Twitter, Facebook, or Instagram.

Customer Experience Audit vs. Customer Satisfaction Survey

Leave a reply

Why They Measure Different Things

Customer Experience Audit vs. Customer Satisfaction Survey

by Braden Kelley and Art Inteligencia

“We already survey our customers” is the single most common objection I hear when I raise the idea of an experience audit, and it’s a reasonable one on the surface — why pay for a second measurement of the same thing? The honest answer is that a survey and an audit aren’t measuring the same thing at all. They’re not even measuring in the same direction.

A survey measures what customers are willing to tell you

NPS, CSAT, and CES all share a structural feature: they depend entirely on a customer choosing to respond, and then choosing to be candid in that response. That’s not a flaw in the instrument — it’s simply what the instrument is. It tells you the sentiment of your most engaged customers (response rates skew toward people who feel strongly, in either direction) at a single moment, about the parts of the experience they happened to be thinking about when the survey arrived.

What it structurally cannot tell you: what happened to the customer who didn’t respond. What the friction actually looked like, step by step, that produced a “6” instead of a “9.” Whether a “9” from one customer and a “9” from another represent the same underlying experience, or two very different ones that both happened to land on the same number.

An audit measures what’s actually happening in the journey

An audit doesn’t ask customers to self-report — it walks the journey directly, the way a real customer experiences it, and documents what’s actually there. That distinction matters most exactly where surveys go quiet: the steps a customer takes for granted and never thinks to mention, the workaround they built without realizing it was a workaround, the moment where the process technically succeeded but took four times longer than it should have.

You can walk journeys for clients whose NPS had been flat, technically acceptable, for years — and find friction serious enough to explain real revenue loss, sitting in a step that had simply never come up in a survey response because no customer thought to complain about something they’d quietly adapted to.

Where each one actually earns its place

None of this makes the survey obsolete — it makes it a different tool for a different job. A survey is the right instrument for tracking sentiment trend over time, cheaply and continuously, across your whole customer base. It’s the wrong instrument for finding out why the trend is what it is, or for finding the friction nobody thought to mention.

An audit is the right instrument for that “why” — for producing a specific, prioritized map of where the experience actually breaks, ranked by business impact. It’s not something you run monthly; it’s something you run when the survey data has told you that something’s wrong without telling you what.

The tell that you need one, not the other

If your satisfaction scores have been flat — not declining, just plateaued — despite genuine effort to improve them, that’s usually the clearest signal that the problem lives somewhere the survey can’t see, and that more survey data won’t produce a different answer than the data you already have. That’s the specific situation an audit is built for.

If you want a rough sense of what that plateau might be costing before committing to a diagnosis, the CX ROI Calculator is a fast way to put a number on it. When you’re ready to find out exactly where the friction is living, that’s what a Customer Experience Audit is for.

Customer Experience Audit versus Customer Satisfaction Survey Infographic

Want to learn more about the value of having an independent Customer Experience Audit done? Or are you ready to invest in one?

Image Credit: Gemini, ChatGPT

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Claude to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Building the Business Case for a Customer Experience Audit

Leave a reply

What the C-Suite Actually Asks

Building the Business Case for a Customer Experience Audit

by Braden Kelley and Art Inteligencia

Every Customer Experience (CX) leader I’ve worked with believes, correctly, that their organization needs a customer experience audit. Very few of them get the budget approved on the first try. The gap almost never comes down to whether the need is real — it comes down to whether the person championing it walked into the room prepared for the four questions a C-suite reliably asks, in roughly this order.

“What does this cost us if we do nothing?”

This is the opening question, and it’s the one the CX ROI Calculator exists to answer. Walk in with your own churn rate, revenue per customer, and a modeled range — conservative to optimistic — rather than an industry statistic borrowed from a research report. A number that’s obviously yours survives scrutiny. A number that’s obviously generic invites the room to argue with the source instead of the substance.

“Why an audit, and not just another survey?”

This is where most business cases quietly fall apart, because the honest answer requires admitting a limitation of what you’re already doing. Your NPS and CSAT programs measure what customers are willing to tell you. An audit measures what’s actually happening in the journey, including the parts customers work around instead of reporting. If your organization has been running satisfaction surveys for years and CX metrics still haven’t moved the way they should, that’s not evidence the audit is unnecessary — it’s usually the single best evidence that it is. Surveys have had their chance to find the problem. They haven’t. A different method is the correct next step, not a redundant one.

“What will we actually be able to do differently afterward?”

An executive approving a budget is not funding a diagnosis for its own sake — they’re funding the decisions the diagnosis will enable. The strongest version of this answer is specific: an audit produces a prioritized list of friction points ranked by business impact, not a general health score. Walk in already able to name the kind of decision it unlocks — “we’ll know whether to fix onboarding or billing first” is a far stronger sentence than “we’ll understand our customers better.”

“How disruptive is this, and how long until we see something?”

This is the question that kills otherwise-approved initiatives at the last step, usually because nobody addressed it until it was asked live in the room. Have the realistic timeline ready before you’re asked for it, not after: when the audit starts, what it requires from internal teams, and when the first findings arrive. Vagueness here reads as risk, even when the actual answer would have been reassuring.

Sequencing the case correctly

The order matters as much as the content. Lead with the cost of inaction (the number), and the room is primed to hear the diagnosis as the obvious next step rather than an added expense. Lead with the audit itself, and you’re immediately negotiating from a weaker position — explaining a cost before anyone in the room has agreed there’s a problem worth solving.

If you haven’t run your own numbers yet, start with the calculator — it’s the fastest way to walk into that first conversation with your own defensible figure instead of someone else’s. When you’re ready to talk about what an audit specifically finds and how it runs, the audit page has the detail, and I’m glad to answer the disruption and timeline questions directly if you’d rather hear them from me before you’re asked them by your own leadership.

Building the Business Case for a Customer Experience Audit

Image Credit: Gemini, ChatGPT

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Claude to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

The Customer Experience Costs Your ROI Calculator Can’t See

Leave a reply

The Customer Experience Costs Your ROI Calculator Can't See

by Braden Kelley and Art Inteligencia

If you’ve run your numbers through the CX ROI Calculator, you already have a real, defensible number — built on your churn rate, your revenue per customer, and the same research-backed value chain I’ve written about before. That number is useful. It’s also almost certainly an undercount, and it’s worth understanding exactly why before you present it as the whole picture.

The model only sees what you’re already measuring

The four-step value chain — metric moves, behavior changes, revenue follows — is a genuinely good way to translate NPS or CSAT into dollars. But notice what it depends on: an experience metric you’re already tracking. That’s the model’s strength and its blind spot in the same breath. It can only quantify the friction that shows up in a score someone bothered to give you.

Four Step Value Chain

Most friction doesn’t show up in a score. It shows up nowhere, until it shows up in the renewal number six months later.

Three costs that live outside the metrics

The silent downgrade. A customer who’s frustrated rarely cancels immediately. More often, they quietly reduce usage, delay an upgrade they were considering, or let a seat go unfilled at renewal instead of adding the three they’d planned to add. None of that trips a churn alert — churn alerts fire on cancellation, not on quiet contraction. By the time it’s visible in a churn dashboard, you’re measuring the outcome of a decision the customer made months earlier, for reasons nobody on your team ever heard about.

The workaround. When something in the experience is broken, customers don’t reliably tell you — they build a workaround and keep using your product anyway. I’ve sat in on customer interviews where someone described, almost proudly, a twelve-step manual process they’d built to avoid a feature that didn’t work the way they needed. That customer will show up in your NPS survey as a “7” — not a detractor, not a promoter, just quietly tolerating a cost you don’t know exists. A workaround is a real cost to serve, it just never gets coded as a support ticket or a complaint.

The frontline save. Your support and success teams are, right now, absorbing friction on your behalf — smoothing over a confusing invoice, manually fixing what an automated process got wrong, apologizing for something they didn’t cause. Every one of those saves is a real cost (in time, in morale, in the eventual departure of your best frontline people), and every one of them is specifically designed, by the person doing it, to be invisible to leadership. That’s their job. It also means your dashboards are structurally blind to exactly the problems your best people are working hardest to hide from you.

Why this isn’t an argument against the calculator

None of this is a reason to skip the ROI modeling — a defensible number beats no number, and if you haven’t run yours yet, start there. It’s a reason to be honest about what the number represents: a floor, not a ceiling. It quantifies the experience gaps you can already see. It has no way to quantify the ones nobody’s told you about yet.

That’s the specific gap a Customer Experience Audit is built to close. Where the ROI model starts from your existing metrics and works outward, an audit starts from the actual customer journey — walked directly, not inferred from a survey response rate — and finds the workarounds, the silent downgrades, and the frontline saves before they’ve had time to show up as a number at all. (If terms like “cost to serve” or “revenue leakage” aren’t consistent vocabulary across your team yet, the Experience Design Glossary is a quick way to get everyone aligned before that conversation.)

Run the calculator first. It’ll tell you the size of the problem you already know about. The audit tells you what else is there.

Get the CX ROI Benchmark Report — the full industry benchmark table with sources, the CX Value Chain framework, and answers to the five objections a CFO is most likely to raise. Enter your email and we’ll send it straight to your inbox.

If after exploring the ROI calculator you would like to explore unlocking revenue opportunities for your business with a Customer Experience Audit, contact me directly. I’m happy to have a no-obligation conversation about whether an audit makes sense for your current situation.

Image Credit: Gemini

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Claude to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Why CSAT Can Look Fine While Revenue Leaks

Leave a reply

Why CSAT Can Look Fine While Revenue Leaks

by Braden Kelley

Your dashboard says customers are satisfied. CSAT is green. The quarterly deck gets a polite nod. Meanwhile, expansion stalls, renewals quietly soften, and support costs creep up — and nobody can point to a single “bad” survey score that explains it.

That isn’t a mystery. It’s a metric paradox: CSAT can look fine while revenue leaks, because satisfaction surveys and financial outcomes measure different things on different clocks.

If you lead CX, product, support, or a P&L, this gap is where budget conversations die. Leaders feel the leak. The score refuses to confess. So the investment stalls.

What CSAT Actually Measures (and What It Doesn’t)

CSAT usually answers a narrow question: how satisfied was someone with a specific interaction or recent experience? That’s useful. It is also incomplete.

CSAT tends to miss:

  • Silent churn — customers who never complain, then don’t renew, don’t expand, or quietly reduce usage
  • Effort and friction — people who “succeed” after three workarounds and still tick Satisfied because the alternative was worse
  • Non-respondents — the angry and the indifferent often skip the survey; the polite remain
  • Journey seams — handoffs between marketing, sales, onboarding, billing, and support where trust dies between touchpoints
  • Lag — revenue damage compounds for months before it shows up as a churn spike leadership will fund

So the score can stay “fine” while the experience failures draining your P&L keep working.

The Metric Paradox in Plain Language

Here’s the pattern I see in Customer Experience Audits:

  1. A customer hits friction (confusing onboarding, surprise fees, repeated authentication, a broken promise after purchase).
  2. They still complete the task — eventually — so the transactional CSAT looks acceptable.
  3. They tell fewer colleagues. They stop exploring add-ons. They price-shop next renewal. They open more tickets.
  4. Finance sees softer expansion and higher cost-to-serve. CX sees a green dashboard.
  5. Both sides are “right” inside their metrics — and wrong about the business.

CSAT is not lying. It’s answering a different question than the one the CFO is asking.

Why Leaders Trust the Wrong Green Light

Organizations over-index on CSAT (and sometimes NPS) because the number is:

  • Familiar in the board pack
  • Easy to benchmark
  • Simple to own in a slide

What’s missing is the chain from experience → behavior → dollars. Without that chain, “improve CX” sounds like a vibe. With it, friction becomes a funding conversation.

I’ve written separately about how to calculate customer experience ROI using that chain. This piece is about why you need it even when — especially when — CSAT looks fine.

Five Signs CSAT Is Masking Revenue Leakage

  1. High CSAT, flat or falling expansion — satisfied enough to stay, not inspired to buy more.
  2. High CSAT, rising contact rate — people are “satisfied” with heroic recoveries you shouldn’t need.
  3. High CSAT in support, weak onboarding completion — you’re measuring the rescue, not the journey.
  4. Promoters who still churn on price — affection without switching costs or realized value.
  5. Teams arguing about the score instead of walking the journey — the map has replaced the territory.

If two or more of these feel familiar, your dashboard is under-reporting risk.

What to Measure Alongside CSAT

Keep CSAT. Add instruments that speak to money and effort:

  • Leading behaviors: activation, time-to-value, repeat purchase, expansion, referral attempts
  • Effort: CES or task completion without assistance
  • Cost-to-serve: contacts per customer, repeat contacts, escalation rate
  • Experience Level Measures (XLMs): human-success metrics tied to specific “ugh” moments — not just uptime SLAs (more on XLMs here)
  • Journey evidence: what an outside-in audit finds when someone actually walks the experience

Scores without journeys produce false calm. Journeys without dollars produce false urgency. You need both.

Put a Number on the Leak (Even a Conservative One)

You don’t need false precision. You need a credible range that makes the paradox discussable in a budget meeting.

Start with what you already know — customers, revenue per customer, churn, service cost — and estimate what a realistic improvement in retention or cost-to-serve is worth annually.

Customer Experience ROI Calculator

Use the free Customer Experience ROI Calculator →

It runs that estimate with your numbers (or industry starting points), and you can copy a summary for a slide. The point isn’t to worship the model. The point is to stop pretending a green CSAT tile equals a healthy P&L.

From Estimate to Action

Once you have a number, the next question is where the leak lives. That’s what a human-centered Customer Experience Audit is for: walk the real journey, find the friction inventory, and prioritize fixes by revenue impact — not by whoever shouted loudest in the last QBR.

CSAT can look fine while revenue leaks. The organizations that pull ahead are the ones willing to measure the leak — then fix the experience that caused it.

Next step: Run the CX ROI Calculator (about two minutes). If the estimate bothers you, that’s useful information — and a good reason to talk about an audit.

The fastest way to see this framework in action is to run it against your own business — enter your customer count, revenue per customer, and current churn rate (or start from an industry benchmark), and it estimates the annual revenue and cost-to-serve impact of a defined experience improvement, along with a summary you can paste straight into a slide.

Get the CX ROI Benchmark Report — the full industry benchmark table with sources, the CX Value Chain framework, and answers to the five objections a CFO is most likely to raise. Enter your email and we’ll send it straight to your inbox.

If after exploring the ROI calculator you would like to explore unlocking revenue opportunities for your business with a Customer Experience Audit, contact me directly. I’m happy to have a no-obligation conversation about whether an audit makes sense for your current situation.

Image Credit: Cursor

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Cursor to clean up the article and add images.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Why Conversations Are the New Digital Gold

Leave a reply

The Big New Revenue Opportunity for Google, OpenAI and Anthropic

Why Conversations Are the New Digital Gold

by Braden Kelley and Art Inteligencia

I. Introduction: The Disruption of the Clickstream

For over two decades, the digital economy operated on a straightforward, predictable currency: the clickstream. Organizations built vast marketing engines, customer experience frameworks, and product strategy backlogs around keyword volumes, cost-per-click (CPC) bidding, and web analytics. If you could capture user search intent at the top of the funnel and guide them through a sequence of web pages, you owned the customer relationship.

That paradigm is experiencing an irreversible structural breakdown. We are witnessing a profound behavioral migration away from typing fragmented queries into a text box toward engaging in fluid, multi-turn dialogue with generative AI assistants. Whether users are speaking directly to Gemini on Android and iOS devices or consulting ChatGPT and Claude for complex decision-making, the mechanics of discovery have fundamentally changed.

The Death of “10 Blue Links”

The traditional search results page — dominated by ranked links, banner inventory, and sponsored listings — is giving way to synthesized, conversational answers. When users speak to an ambient assistant, they aren’t looking for a list of websites to evaluate independently; they are seeking a resolved outcome. Speech-to-text, natural voice interaction, and inline AI reasoning mean problem-solving happens within the dialogue itself, drastically reducing the need to visit external brand properties.

The Shrinking Digital Surface Area

This rise in zero-click interactions presents an existential challenge for traditional web analytics and performance marketing. As consumer click-through rates decline, brands face a dramatic reduction in their visible digital touchpoints:

  • Attribution Blindness: Traditional conversion tracking breaks down when the research and evaluation phases occur entirely inside an AI model’s context window.
  • Diminishing SEO Returns: Optimizing for keywords and site traffic yields shrinking returns when AI models synthesize answers directly without referring users to source URLs.
  • Loss of Direct Engagement: The digital surface area where brands can present their unique visual identity, messaging, and experience design is rapidly compressing.

The Foresight Premise

In any major technology transition, structural shifts create immediate information asymmetries. Every change initiative produces winners and losers based on who recognizes where value is re-aggregating. The primary battleground of the AI era is no longer about driving traffic to a destination — it is about controlling, understanding, and translating the rich context of human conversational intent.

II. The Blind Spot: How Brands Are Losing the Voice of the Customer

The transition from traditional web search to ambient AI interaction is creating an unprecedented intelligence blackout for commercial enterprises. For years, organizations refined their understanding of consumer behavior by tracking the digital breadcrumbs left across search engines, landing pages, and digital storefronts. As customer decision-making migrates into private, dynamic AI dialogues, that pipeline of actionable data is drying up.

This shift represents far more than a marketing disruption — it is a fundamental erosion of the qualitative feedback loops that drive modern product innovation and experience design.

From Keywords to Unfiltered Intent

Keyword search was always a compromised, low-fidelity medium. Users learned to compress their complex human needs into unnatural, fragmented phrases meant to nudge a search algorithm into producing useful links. The language of traditional search was structured around constraints rather than context.

Generative AI and voice interfaces have eliminated those constraints. When individuals speak to an assistant like Gemini, ChatGPT, or Claude, they express their needs with full nuance, nuance, and emotional framing. Consider the structural difference between these two modes of inquiry:

  • Traditional Search Query: best running shoes flat feet
  • Conversational Intent: “I’m training for my first rainy marathon in three months, but I have mild overpronation and a old knee injury. What shoes under $150 will give me enough stability without causing blisters on long runs?”

The conversational prompt contains rich layers of context: budget parameters, timeline constraints, physical vulnerabilities, weather considerations, and personal goals. However, because this interaction takes place within an AI context window rather than on a brand’s website or an open search results page, the business whose product is being evaluated receives zero visibility into the exchange.

The Customer Insight Vacuum

As consumer preference formation moves into continuous multi-turn conversations, brands are losing access to critical moments of truth across the buyer journey. This creates three severe operational blind spots:

  • Unseen Feature Trade-offs: Brands cannot see which specific product attributes, specifications, or pricing structures cause a potential customer to eliminate them from consideration during an AI dialogue.
  • Invisible Competitive Comparisons: When an AI assistant evaluates three competing solutions side-by-side for a user, the losing brands receive no signal explaining why the model recommended an alternative.
  • Obsolete Voice-of-Customer (VoC) Data: Traditional surveys, focus groups, and social listening tools capture lagging, highly filtered opinions. They fail to reflect the real-time, unvarnished friction points articulated during natural conversations with AI.

The Experience Design Risk

Without access to the rich contextual signals embedded in everyday user prompts, corporate experience design initiatives risk operating on outdated assumptions. Customer journey maps, persona frameworks, and friction-point analyses quickly become stagnant snapshots of an obsolete digital funnel.

To design meaningful, human-centered experiences, leaders must understand the authentic language and evolving expectations of their audience. When that language is spoken exclusively to third-party AI assistants, organizations that fail to secure access to conversational intelligence will find themselves innovating in the dark.

III. The Big Pivot: Monetizing Context, Not Clicks

Every major shift in technology redistributes economic value. As traditional cost-per-click advertising yields diminish under the pressure of zero-click conversational answers, the business models of the AI platform giants — Google, OpenAI, and Anthropic — must evolve. The next multi-billion-dollar monetization opportunity will not come from placing banner ads inside conversation flows, but from harvesting, structuring, and licensing the vast reservoir of real-time human intent being shared with their models every second.

Human conversation is the new digital gold. For businesses desperate to recover lost visibility into the buyer journey, aggregated conversational intelligence represents the ultimate strategic asset.

The New Revenue Engine for AI Titans

Advertising models built on static keyword triggers are fundamentally mismatched with fluid, multi-turn AI reasoning. Forcing intrusive sponsored links into a personalized voice response destroys the user experience. Instead, AI providers are positioned to monetize the output side of their platforms by acting as enterprise data brokers, transforming raw dialogue logs into high-value intelligence feeds.

By capturing how millions of people naturally discuss needs, compare options, and express frustrations, platform owners can package anonymized context into enterprise-grade analytics products that command recurring software-as-a-service (SaaS) subscription premiums.

Packaging the “Digital Gold”

This new intelligence layer will yield actionable commercial products tailored for product strategists, marketers, and executive leaders:

  • Brand Health & Recommendation Telemetry: Real-time quantitative dashboards tracking how frequently a brand is mentioned during advice seeking, the sentiment surrounding those mentions, and the exact contexts in which competitors are favored.
  • Unmet Need & Latent Demand Mapping: Algorithmic extraction of emerging consumer pain points long before they manifest in formal search trends, support tickets, or market research reports.
  • Decision Boundary & Friction Analysis: Synthesized reports detailing the specific trade-offs (price points, missing features, usability concerns) that systematically cause prospective buyers to reject a product during AI-driven evaluations.

Democratizing Enterprise Intelligence

The power of conversational analytics lies in its scalability across the economic spectrum. While enterprise corporations will pay premium tiers for custom API integrations and real-time category alerts, small and medium-sized businesses (SMBs) will finally gain access to market research previously reserved for Fortune 500 budgets.

A local bike shop or boutique software firm could subscribe to a regional category feed to instantly discover the precise features or price barriers driving customer choices in their specific niche. By turning unvarnished human dialogue into structured insight, AI platforms will unlock an indispensable revenue model powered by authentic human context.

IV. Human-Centered Change & Ethical Governance

Unlocking the commercial value of conversational data requires navigating a complex intersection of consumer trust, regulatory compliance, and organizational transformation. Because natural language dialogue contains deep personal context, commercializing this information demands rigorous ethical boundaries. The success of conversational intelligence as a revenue model hinges on maintaining strict user privacy while helping enterprises build the internal capabilities needed to act on these new insights.

Privacy by Design: The Ethical Imperative

Monetizing conversational context cannot come at the expense of individual privacy. AI platform operators must engineer robust data architecture standards that prevent the exposure of personally identifiable information (PII) while preserving strategic utility:

  • Differential Privacy & Aggregation: Injecting mathematical noise into datasets so macro-level consumer trends can be analyzed without ever exposing individual user transcripts.
  • Synthetic Data Modeling: Generating artificial, representative datasets derived from real conversation patterns, allowing brands to analyze buyer behavior without touching live user interactions.
  • Strict Brand-Level Anonymization: Ensuring that enterprise dashboards expose category-level intent and competitive positioning without revealing specific user identities or sensitive personal attributes.

Overcoming the “Surveillance” Backlash

Public perception will determine the speed at which conversational analytics becomes mainstream. If consumers view the monetization of their conversations as invasive surveillance, user churn and regulatory pushback will quickly follow. AI providers and brands must collectively frame conversational analytics around mutual value creation.

When customer intent data is anonymized and applied ethically, it leads directly to better product design, more intuitive user interfaces, and the elimination of persistent market friction points. The objective must be presented clearly: using collective, human-centered feedback to build products and experiences that better serve actual human needs.

Managing Organizational Readiness

Accessing conversational intelligence is only half the equation; corporate leadership teams must also transform how they make decisions. Applying the principles of Human-Centered Change™, organizations must actively prepare their cultures, workflows, and talent to interpret fluid conversational data rather than static web metrics.

This operational transition requires shifting leadership focus away from legacy digital KPIs like bounce rates, page views, and click-through rates toward modern conversational indicators: share of voice in model recommendations, prompt inclusion rates, and conversational intent fulfillment. Companies that successfully align their internal culture around these human-centered insights will build an enduring competitive advantage in the AI era.

V. FutureHacking™: Strategic Implications for Business Leaders

To navigate the shift from transactional clickstreams to continuous conversational context, executive leadership cannot afford a reactive stance. Applying a FutureHacking™ lens — scanning weak signals around emerging user behaviors today to anticipate the structural realities of tomorrow — reveals a multi-phase transformation in how organizations will make decisions, design experiences, and compete for market share.

The transition toward conversational intelligence will unfold across three distinct horizons over the next decade.

Near-Term Horizon (1–2 Years): The Rise of Generative Engine Optimization & Intelligence Pilots

In the immediate term, traditional Search Engine Optimization (SEO) will yield ground to Generative Engine Optimization (GEO). As organic web traffic declines, brands will pivot from optimizing page headers and backlinks to structuring brand narratives and product specifications so they are accurately ingested and cited by foundational AI models.

Concurrently, early adopter enterprises will join private pilot programs hosted by Google, OpenAI, and Anthropic. These initial telemetry dashboards will give brand managers their first high-level visibility into prompt inclusion rates, category mention frequencies, and overall model recommendation sentiment.

Medium-Term Horizon (3–5 Years): Synthetic Focus Groups & Simulated Customer Journeys

As the granularity of anonymized conversational datasets improves, market research will undergo a radical evolution. Rather than waiting weeks to conduct traditional focus groups or analyze retrospective survey results, product strategy teams will query specialized AI models trained on billions of real-world conversational signals.

Organizations will routinely run product concepts, pricing adjustments, and brand positioning messaging against synthetic persona populations. These simulated customer panels will instantly predict friction points, feature trade-offs, and competitive migration risks based on real-time consumer intent trends, drastically compressing product development cycles.

Long-Term Horizon (5+ Years): Closed-Loop Innovation Systems

Over a five-year horizon, conversational intelligence will move from a passive diagnostic tool to an active driver of automated organizational workflows. Leading enterprises will construct closed-loop innovation engines where real-time conversational data directly informs cross-functional operations:

  • Automated Backlog Prioritization: Product engineering roadmaps will dynamically re-prioritize feature requests based on unprompted feature complaints captured across category-wide AI dialogues.
  • Dynamic Experience Adaptation: Digital touchpoints and customer service flows will auto-tune their messaging and support options based on emerging friction patterns identified by ambient assistants.
  • Continuous Portfolio Alignment: Mergers, acquisitions, and line extensions will be evaluated using continuous, real-time demand signals extracted directly from human-AI problem-solving sessions.

By anticipating these structural horizons today, forward-thinking leaders can begin building the data infrastructure, talent capabilities, and agile decision-making frameworks required to turn conversational signals into market leadership.

VI. Conclusion & Key Takeaways for Innovators

The transition from transactional keyword search to ambient, multi-turn AI dialogue represents one of the most profound structural shifts in the history of the digital economy. As consumers speak directly with Gemini, ChatGPT, and Claude on their mobile devices and desktop interfaces, the clickstream era is drawing to a close. Waiting for traditional web traffic, cost-per-click efficiency, and search ad impressions to recover is not just an ineffective strategy — it is an existential risk.

The organizations that thrive in this next era will be those that recognize where strategic value has re-aggregated: away from driving website visits and toward capturing, understanding, and acting upon authentic conversational context.

Key Takeaways for Business Leaders

  • Acknowledge the Intelligence Blackout: Traditional SEO, web analytics, and click-through attribution models are providing a rapidly shrinking window into true customer behavior. Accepting this loss of visibility is the first step toward building modern, conversation-aware capabilities.
  • Prepare for the Conversational Data Economy: As traditional search advertising revenues face long-term pressure, Google, OpenAI, and Anthropic will monetize anonymized conversational data. Forward-thinking leaders should allocate budget now for emerging conversational telemetry feeds and Generative Engine Optimization (GEO).
  • Embed Human-Centered Change™: Shifting an organization from static KPIs (page views, bounce rates) to conversational metrics (share of voice in model answers, prompt inclusion, intent fulfillment) requires intentional change management. Re-align leadership, cross-functional teams, and innovation pipelines around these new signals.
  • Rethink Experience Design: Continuous multi-turn dialogues reveal unvarnished human friction points, budget constraints, and feature trade-offs. Integrate these real-time qualitative signals into your customer journey maps and product development roadmaps to eliminate customer friction faster than competitors. Invest in a Customer Experience Audit to find where you fall short.

Data was the primary oil of the early web era, but synthesized human conversation is the true gold of the AI era. By pairing ethical governance and human-centered design with the rich intent embedded in everyday dialogue, innovative organizations can illuminate their blind spots, transform their decision-making, and create products that resonate with authentic human needs.

Frequently Asked Questions

Why are traditional search advertising and click-through rates declining?
As users shift from keyword-based search boxes to ambient AI assistants like Google Gemini, ChatGPT, and Claude, they receive direct, synthesized answers rather than a list of web links. This rise in zero-click interactions significantly reduces website referral traffic and traditional ad impression volume.
How do AI platforms like Google, OpenAI, and Anthropic plan to monetize conversational data?
AI platform providers can package anonymized, aggregated conversation logs into enterprise intelligence feeds. By selling brand health telemetry, unmet need analytics, and consumer friction insights to businesses, AI companies create a massive new recurring revenue stream to complement or offset declining search ad yields.
How can businesses prepare for the shift from keyword search to conversational intelligence?
Organizations must transition their digital strategy from traditional SEO to Generative Engine Optimization (GEO), adapt internal change management frameworks (such as Human-Centered Change™) to track conversational metrics like model share-of-voice, and subscribe to emerging conversational analytics feeds to inform product design and experience strategies.


Image credits: Gemini

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Gemini to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Surveys Are Collapsing

Leave a reply

Conversational and Agentic VoC is How Loyalty Gets Heard

Conversational and Agentic VoC is How Loyalty Gets Heard

by Braden Kelley and Art Inteligencia

The Quiet Collapse of the Survey Layer

Something uncomfortable is happening inside customer experience programs that still treat the survey as the source of truth. Response rates are falling — sometimes sharply — even when the questionnaire itself barely changes. The invitations still go out. The dashboards still refresh. The air getting thinner is the percentage of customers willing to talk to a form.

This is not the death of listening. It is the collapse of a layer: the assumption that loyalty, satisfaction, and experience quality can be reliably extracted on demand through static instruments. Net Promoter Score is not vanishing overnight. Forms are not obsolete tomorrow morning. But both are being demoted — from verdict to signal, from system of record to starting point.

Organizations that built governance, bonuses, and “voice of the customer” theater almost entirely on survey completion are discovering a hard truth of human-centered change: when the method stops matching how people communicate, the method stops producing wisdom. You can still report a number. You just cannot pretend it represents the relationship.

The urgent question for innovators is not how to squeeze three more points of response rate out of a dying habit. It is how to hear customers in the ways they already speak — and how to turn that listening into action before loyalty quietly leaves.

Why People Stopped Talking to Forms

People did not become less opinionated. They became less willing to perform unpaid labor for brands that ask without reciprocating.

Survey fatigue is real, but it is only the surface. Timing is often wrong — a form arrives after the emotional moment has passed, or in the middle of a busy day when the only honest answer is delete. Reciprocity is weak: customers complete the ritual and see no change, so the next invitation feels like noise. Channel mismatch is growing: people already live in chat, voice, messaging, and short conversational bursts, while VoC programs still insist on a clipboard with radio buttons.

Underneath the mechanics sits an emotional job. Feedback, at its best, is a bid to feel heard. A form rarely delivers that feeling. It flattens story into score, urgency into scale, and dignity into “additional comments (optional).” When the experience of giving feedback is itself a poor experience, silence becomes rational.

Human-centered leaders should treat declining response as diagnostic data. Customers are telling you — by not answering — that your listening design is out of date.

From Scorekeeping to Sense-Making

Traditional VoC optimized for scorekeeping: capture a metric, trend it, threshold it, celebrate or panic. Sense-making asks a different question: What is changing in the lived experience, and why?

In a post-survey-dominant world, unstructured signal matters more — conversations, call notes, chat transcripts, reviews, social fragments, support themes, behavioral break points. AI makes synthesis of that mess newly practical. That does not make the score useless. It makes idolatry of the score dangerous.

The “why” can no longer be an afterthought parked in an open text field that nobody has time to read. The why is the product of modern listening. Scores become navigation lights. Narratives, patterns, and emotions become the map.

This shift also changes operating rhythm. Quarterly report theater gives way to continuous closed loops: hear, understand, act, confirm. Loyalty intelligence is less a research project and more an always-on sense-making system — still human-governed, still ethically bounded, but finally matched to the speed at which experience actually breaks.

Conversational VoC: Feedback as Dialogue

Conversational VoC replaces the clipboard with a dialogue. Instead of forcing every customer through the same static path, listening adapts — in the moment, in the channel, and in response to what the person just said.

That can look like a short adaptive chat after a key journey step, a voice interview that follows curiosity instead of a rigid script, a messaging thread that asks one good question and then the next logical one, or a human interview amplified by better prompts and synthesis. The common design principle is simple: treat feedback as conversation, not compliance.

Dialogue earns what forms forfeit. It can hold emotion without collapsing it into a single digit. It can clarify ambiguity in real time. It can meet people where they already are speaking. And it can make reciprocity visible — “we heard you, here is what happens next” — which is how listening becomes trust rather than extraction.

Done poorly, conversational VoC is just a survey wearing a chatbot costume. Done well, it is experience design applied to insight itself: respectful of time, responsive to context, and worthy of the story a customer is willing to share.

Agentic Listening: When Insight Can Act

The next leap is agentic listening: systems that do not only collect and classify, but can route, summarize, prioritize, trigger recovery, and help close the loop across teams. Insight stops dying in a dashboard and starts moving work.

This is powerful — and easy to get wrong. An agent that escalates a frustrated customer to a human with full context is care at scale. An agent that silently profiles, nudges, or “manages” sentiment without consent is surveillance with a CX badge. Human-centered innovation draws that line in the architecture, not in the press release.

Design stakes for agentic VoC

  • Consent and clarity — people should understand when listening is active and how their words will be used.
  • Privacy and minimization — collect what you need for learning and recovery, not everything you can.
  • Escalation with dignity — automation should accelerate help, not trap emotion in a loop.
  • Action accountability — if the system can trigger work, someone must own whether that work actually improved the experience.

Agentic VoC is not a replacement for human judgment. It is orchestration for listening: machines handle volume and routing; people handle meaning, ethics, and relationship repair. The brands that win will be the ones whose listening systems can act — and whose customers still feel respected while they do.

A Human-Centered Playbook for the Post-Survey Era

You do not need to burn the survey. You need to dethrone it. Here is a practical path.

  • Keep scores as signals, not idols. Use them to notice change; use conversations and behavior to explain it.
  • Build conversational intake at moments that matter. Short, adaptive, channel-native dialogues beat long retrospective forms.
  • Unify experience data. Connect feedback, journeys, and operational reality so insight is not stranded in a research silo.
  • Close loops where customers can feel them. Private recovery for individuals; visible improvement for patterns. Reciprocity is the antidote to silence.
  • Measure whether people feel heard — and whether action followed. Listening quality is an experience metric, not only a research metric.
  • Govern agentic listening for care. Decision rights, consent, escalation, and audit trails before autonomy scales.

Futurology in customer experience is often sold as more instrumentation. The deeper shift is more humane instrumentation: listening that fits human communication, sense-making that honors story, and systems that can act without making people feel managed.

Surveys are collapsing as the center of gravity. Conversational and agentic VoC are how loyalty gets heard again — not as a quarterly score, but as a living relationship that organizations are finally designed to understand.

Frequently Asked Questions

Why are customer survey response rates declining?

Response rates are falling because of survey fatigue, poor timing, weak reciprocity when feedback leads to no visible change, and a mismatch with how people already communicate through chat, voice, and messaging. Many customers still have opinions — they are less willing to share them through static forms.

What is conversational VoC?

Conversational voice of the customer (VoC) gathers feedback through adaptive dialogue — such as chat, voice, or messaging — rather than fixed questionnaires. It follows context and emotion in the moment, making customers more likely to feel heard and producing richer insight into the why behind experience scores.

What is agentic VoC and how does it differ from surveys?

Agentic VoC uses AI systems that can not only collect and analyze feedback but also route issues, trigger recovery, summarize themes, and help close the loop. Unlike surveys that mainly capture scores after the fact, agentic listening turns insight into action — when governed with consent, privacy, and human escalation.

Image credits: Cursor

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Cursor to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Designing Agentic Customer Experience That Earns Trust

Leave a reply

When AI Agents Act on Your Behalf

Designing Agentic Customer Experience That Earns Trust

by Braden Kelley and Art Inteligencia

From Answers to Actions: The Agentic Shift

For a decade, “AI in customer experience” mostly meant better answers: chatbots that deflected, assistants that summarized, copilots that drafted. Helpful, imperfect, and still largely conversational. The agentic shift is different. Systems are no longer limited to recommending what a human should do. They are beginning to do — refund, reschedule, rebook, reroute, update records, trigger fulfillment, and coordinate multi-step journeys across channels without waiting for a ticket to crawl through three departments.

That is not a feature upgrade. It is a change in the relationship. The brand now includes a non-human actor with authority. When an agent acts, it acts in the company’s name and, increasingly, on the customer’s behalf. Experience design can no longer stop at tone of voice and containment rates. It must account for delegated power.

Human-centered innovators should hear the signal underneath the hype. Customers are not primarily evaluating whether your AI sounds clever. They are evaluating whether your organization is safe to trust with unfinished business. Answering a question poorly is friction. Acting incorrectly — or acting opaquely — is a breach of the emotional contract.

Welcome to agentic customer experience: where loyalty is shaped less by what the brand says, and more by what its agents are allowed to decide.

Why Customers Will Forgive Slowness — But Not Betrayal

Customers have always traded time for confidence. Many will wait for a competent human. Far fewer will repeatedly educate a system that forgets context, loops through the same failed path, or blocks the exit to a person. Research across the industry keeps pointing to the same pattern: openness to AI rises when it resolves issues completely — and collapses quickly when it wastes attempts, hides escalation, or makes people feel trapped.

This is where leaders misread the risk. They optimize for speed and deflection, then wonder why trust erodes. People will often forgive slowness when they feel progress and respect. They will not forgive what feels like betrayal: decisions that seem optimized for the brand’s cost curve, recommendations that ignore stated preferences, silent policy enforcement with no explanation, or “self-service” that is really forced service.

Betrayal in CX is usually quiet. It looks like a denied refund with no rationale. An agent that “helps” by steering toward what is easiest to contain. A personalization engine that remembers everything except the customer’s dignity. The nervous system keeps score. So does the switching decision.

In the agentic era, the question is not only Did we resolve it? It is Did we resolve it in a way that still makes this relationship feel safe?

The New Experience Design Problem: Delegation

Most AI roadmaps are still framed as automation problems: what can we remove from the human queue? That framing is incomplete. From the customer’s side, agentic CX is a delegation problem. People are deciding how much unfinished business they are willing to hand to a system that can act without them in the room.

Delegation requires a different design brief. Customers need to know what is being done, why it is being done, what happens if it goes wrong, and how to reclaim control. Without those conditions, “autonomy” feels like abandonment dressed up as innovation.

Human-centered experience design therefore asks emotional jobs beneath the functional ones:

  • Do I feel represented — or processed?
  • Do I feel informed — or surprised after the fact?
  • Do I feel able to intervene — or locked out by design?
  • Do I feel the brand is on my side — or merely efficient at managing me?

This is why transparency is not a compliance garnish. It is part of the product. So is the handoff. An elegant agent that cannot escalate with context intact is not advanced; it is brittle. Agentic excellence includes knowing when not to act alone.

Four Trust Pillars for Agentic CX

If agentic systems will act in your name, trust needs architecture — not slogans. Four pillars help leaders design for loyalty rather than mere containment.

Clarity

Customers should understand when AI is involved, what it can and cannot do, and what just happened. Clarity reduces suspicion. Mystery breeds it. “Transparency by design” means visible agency, plain-language explanations, and no dark patterns that disguise automation as a person.

Competence

An agent that acts must finish the job. Partial resolution, lost context, and repetitive failure teach customers that delegation is unsafe. Competence is end-to-end: data continuity, accurate policy application, and the ability to complete multi-step work without making the customer re-narrate their life story.

Control

Trust grows when people can undo, override, confirm high-stakes actions, and reach a human without being punished for asking. Control is not the enemy of automation; it is what makes automation acceptable. The best agentic experiences feel powerful and reversible.

Care

The decisive pillar: whose interest is being optimized? If customers believe the agent is steering them toward what is best for the brand — upsell, denial, deflection — loyalty decays even when the interaction is fast. Care means designing decision logic that is fair, explainable, and aligned with the customer’s stated goal.

Clarity, competence, control, and care. Miss one, and agentic CX becomes a trust tax. Honor all four, and autonomy becomes hospitality at scale.

Orchestration Without Losing the Human

The winning model is not AI-only theater. It is orchestration: purposeful sequencing of agentic action, human judgment, and channel continuity so the customer experiences one coherent journey instead of a relay race of disconnected tools.

Agentic AI is uniquely suited to routine multi-step work — the operational choreography that used to create delay and handoff fatigue. Humans remain essential for ambiguity, emotion, ethical judgment, and exceptions that policies cannot pre-chew. CX leaders increasingly expect human interactions to become more complex as AI absorbs the simple. That is not failure of automation. That is the work migrating to where empathy and discernment still matter most.

Orchestration also includes the employee experience. If frontline teams inherit broken context, unexplained agent decisions, and no authority to repair trust, customers will feel that fracture immediately. Human-centered change treats agents and employees as one system: AI handles volume and velocity; people handle meaning and recovery.

Design the sequence, not just the bot. Decide what should happen before, during, and after autonomous action. Make escalation a first-class journey path, not a hidden defeat. In agentic CX, the brand is the conductor. The technology is the orchestra. Customers can tell when nobody is conducting.

A Human-Centered Playbook for the Agentic Era

Urgency without a playbook produces demos. Loyalty requires operating discipline. Start here.

  • Define decision rights before you deploy autonomy. Which actions can an agent take alone, which require confirmation, and which are human-only? Write it as policy customers can feel in the experience.
  • Design recovery as carefully as resolution. Every autonomous action needs an undo path, an explanation path, and a dignified escalation path with context preserved.
  • Measure trust outcomes, not only efficiency. Containment and average handle time matter. So do repeat contact, forced re-explanation, escalation friction, complaint themes, and whether customers say they would delegate again.
  • Prototype agent behavior on real journeys. Test the emotional arc of delegation: consent, action, visibility, completion, and repair. Bodies and language reveal failure faster than dashboards.
  • Govern for care in public. State how data is used, how models decide, and how you prevent brand-first bias. Trust compounds when principles are operational, not ornamental.

The future of customer experience will not be judged by how many agents you launched. It will be judged by what those agents did in your customers’ names — and whether people still felt human while it happened. Brands that treat agentic AI as a cost play will win quarters. Brands that treat it as a trust system will win relationships.

That is the human-centered mandate of the agentic era: give your systems the power to act, and give your customers every reason to believe that power is being used with them, not on them.

Frequently Asked Questions

What is agentic customer experience?

Agentic customer experience is when AI systems can take multi-step actions on behalf of the customer or company — such as refunds, rescheduling, routing, or journey orchestration — rather than only answering questions. It shifts CX from conversation to delegated action, which raises the bar for trust, transparency, and human handoff.

How can brands build trust in AI agents that act for customers?

Build trust through four pillars: clarity about when AI is acting, competence in completing work with context preserved, control through undo and easy human escalation, and care by optimizing for the customer’s interest rather than containment alone. Recovery design matters as much as automation design.

Will human agents still matter in an agentic CX model?

Yes. Agentic AI is best for routine multi-step work, while humans remain essential for complex, emotional, and exceptional cases. The winning model is orchestration: AI and people working as one system, with seamless escalation and shared context so customers never feel abandoned by automation.

Image credits: Gemini

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Cursor to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Innovation or Not — InTruth

Leave a reply

Innovation or Not — InTruth

by Braden Kelley and Art Inteligencia

Section I: The Context — The Friction of Unfiltered Information

We live in an age of hyper-abundant, instantaneous media. Live-streamed political debates, rapid-fire press conferences, corporate town halls, and unscripted interviews broadcast continuously across our screens. Yet, alongside this unprecedented access lies a compounding crisis: information pollution. Misinformation, selective statistics, and outright falsehoods move at the speed of light — frequently outpacing traditional journalism, post-event debriefs, and manual fact-checking by hours or even days.

By the time a correction is published, the narrative has already taken root in the public consciousness. The damage is done, and the systemic cost to societal trust is immense.

The Human Dilemma: Cognitive Load in the Attention Economy

For the average media consumer, keeping up with live information requires an exhausting mental calculus. Viewers are caught between two undesirable options:

  • Passive Acceptance: Consuming content at face value to save energy, thereby absorbing unverified claims and out-of-context assertions.
  • Active Skepticism: Continuously pausing, opening secondary tabs, and manually searching primary sources to verify claims — a process that destroys flow and creates severe cognitive friction.

In an attention economy built around low-effort, lean-back viewing, asking users to perform manual research while watching live video is a fundamental experience design failure. The friction is simply too high.

Enter InTruth: Shrinking the “Truth Latency”

This is where InTruth enters the frame. Operating as a Chrome extension, InTruth aims to bridge the gap between real-time consumption and rigorous verification. By pairing live automated voice-to-text transcription with continuous, real-time database lookups against verified primary sources, InTruth introduces an automated contextual layer directly over any live video stream.

It seeks to collapse the “truth latency” — closing the critical gap between the moment a claim is uttered and the moment a viewer receives verified context. But does having the capability to cross-reference audio in real time make InTruth a true innovation, or merely a powerful technical demo?

Section II: The Innovation Test — Applying the Value Equation

To evaluate whether InTruth is a genuine breakthrough or merely an intriguing technical novelty, we must look beyond raw technology capabilities and apply a core principle of human-centered change:

Innovation = Value Creation × Value Access × Value Translation

Innovation is never just the underlying technology — it is the holistic system that turns potential value into realized human impact. Because this relationship is multiplicative, if any single variable in the equation drops to zero, the total innovation value collapses to zero. Let’s stress-test InTruth across all three dimensions.

1. Value Creation: Closing the “Truth Latency”

Value Creation represents the raw potential utility of a solution. In the case of InTruth, the potential value is immense. By compressing hours of investigative research into split-second automated lookups, it transforms complex, multi-step investigative work into live, contextual intelligence.

Traditional fact-checking suffers from high “truth latency” — the delay between when a misleading claim is broadcast and when verified facts reach the public. InTruth attacks this latency directly, creating value by:

  • Automating the Research Loop: Instantly translating spoken dialogue into structured search queries against primary databases, public records, and vetted archives.
  • Democratizing Verification: Giving everyday viewers access to research capabilities previously restricted to newsrooms and investigative analysts.
  • Neutralizing Misinformation Velocity: intercepting falsehoods at the exact moment of delivery before they can anchor in the viewer’s memory.

2. Value Access: Reducing Cognitive and Experience Friction

Capability without access is useless. Value Access measures how easily a user can tap into the created value within their existing habits and workflows. This is where experience design becomes the ultimate differentiator.

If InTruth forces users to switch tabs, click through complex menus, or read dense paragraphs of text while trying to watch a fast-paced debate, it introduces severe cognitive friction. To succeed on Value Access, the tool must observe key human-centered principles:

  • In-Flow Delivery: Surfacing unobtrusive, glanceable notifications directly inside the existing video player (whether on YouTube, X, or news networks) without forcing context switching.
  • Cognitive Ergonomics: Presenting live receipts using visual micro-cues (such as color-coded trust vectors or short source snippets) that can be processed at a glance without disrupting the viewing experience.
  • Zero Setup Effort: Functioning ambiently in the background so the user gains immediate insight without configuring complex rules or query filters.

3. Value Translation: Overcoming the Trust Paradox

Great capability and seamless access mean nothing if the user doesn’t believe or understand the output. Value Translation is the art of framing information so that it builds trust and inspires meaningful action.

In fact-checking, Value Translation faces a massive hurdle: human confirmation bias and institutional skepticism. Simply labeling a speaker’s statement as “False” often triggers a defensive psychological reaction, causing viewers to reject the tool rather than reconsider the claim. InTruth must solve this trust paradox by:

  • Sourcing Over Assertion: Shifting away from judgmental labels (e.g., “Pants on Fire”) toward objective, primary-source evidence (e.g., “Congressional Record, Vol. 168: Voted YES on Bill X”).
  • Radical Transparency: Making the verification logic clear, showing exact source documents so users feel empowered to make their own judgment rather than being dictated to by an algorithm.
  • Contextual Nuance: Distinguishing between outright fabrication, missing context, or minor statistical rounding, ensuring the audience trusts the nuance and neutrality of the tool.

Section III: Human-Centered Change & Experience Design Lens

Building a powerful real-time fact-checking engine is fundamentally a human experience challenge, not just an algorithmic one. Technology enables capabilities, but human-centered design enables adoption. To transform InTruth from a novel browser extension into an indispensable daily tool, we must examine the behavioral realities of how people consume media, process conflicting information, and react to real-time feedback.

Designing for Cognitive Load: Preserving the Viewing Flow

Watching live video — whether a town hall, press conference, or political debate — is inherently a passive, low-effort experience. Viewers lean back, absorb narratives, and follow emotional cues. By introducing a real-time verification overlay, InTruth asks the brain to switch into an active, analytical processing mode.

If designed poorly, this extra layer leads to immediate cognitive overload and outrage fatigue. Human-centered experience design must protect the user’s focus through careful interface discipline:

  • Glanceable Micro-Interactions: Information must be visual and bite-sized. Complex policy documents should be distilled into clear, single-line contextual highlights with expandable detail for those who want to dig deeper.
  • Temporal Breathing Room: Notifications must not flash continuously across the screen. Applying intelligent thresholding ensures the overlay only triggers when high-confidence discrepancies or critical missing contexts arise.
  • Visual Hierarchy & Ambient Cues: Utilizing peripheral, non-intrusive status indicators allows viewers to remain immersed in the video stream while maintaining awareness of real-time source verification.

The Source Transparency Model: Spectrum vs. Binary Truth

Traditional fact-checking often relies on reductive binary tags: True or False. In human discourse, however, statements rarely fall into neat black-and-white categories. Rhetoric is built on selective framing, outdated statistics, exaggerated figures, and omitted context.

A rigid binary approach degrades user trust and creates friction. InTruth must adopt a Source Transparency Model that reflects nuance across a spectrum of context:

Verified Primary Record  •  Missing Context  •  Outdated Data  •  Unsubstantiated Assertion

By presenting a spectrum rather than a verdict, InTruth respects the user’s intelligence. It transforms the experience from an automated referee telling the audience what to think into a personal research assistant providing the receipts needed to draw independent conclusions.

Overcoming Psychological Defense Mechanisms

When people are presented with facts that contradict their deeply held beliefs, their immediate instinct is rarely acceptance — it is defensiveness. Cognitive dissonance kicks in, and the brain’s immune system seeks reasons to discredit the source, the algorithm, or the platform delivering the correction.

To navigate this psychological hurdle, InTruth must incorporate key behavioral design tactics:

  • Neutral, Non-Judgmental Language: Avoid emotionally charged terms like “Debunked” or “Lies.” Instead, use objective phrasing such as “Official Treasury data shows…” or “Voting records reflect…”
  • Direct Primary Links: Build instant credibility by linking directly to raw, unedited source files — such as legislative bills, economic reports, or court transcripts — rather than third-party commentary.
  • Empowering User Autonomy: Frame corrections as optional context layers that enrich understanding, ensuring users feel empowered rather than challenged.

Section IV: The Futurology Perspective — FutureHacking™ Live Truth

To understand the long-term strategic trajectory of InTruth, we must look beyond its current implementation as a desktop browser extension. By applying a FutureHacking™ framework — scanning weak signals in technological adoption and anticipating systemic shifts — we can map how real-time truth engines will re-architect the broader media landscape over the next 5 to 10 years.

From Weak Signals to Mainstream Realities

Today, real-time fact-checking overlayed on web video is a weak signal — a niche capability used primarily by journalists, policy analysts, and tech-savvy early adopters. However, several converging trends will accelerate this capability into a baseline expectation for ambient intelligence:

  • The Explosion of Synthetic and AI-Generated Media: As deepfakes, automated audio cloning, and AI-driven propaganda proliferate, unassisted human perception will no longer be sufficient to determine authenticity. Real-time verification will evolve from an optional feature into an essential cognitive defense layer.
  • Ultra-Low-Latency Edge AI: On-device AI models and localized vector databases will allow voice transcription, semantic analysis, and cross-referencing to occur locally with zero network latency, making verification instantaneous and completely private.
  • The Shift to Conversational and Live Streams: As traditional written journalism continues to give way to long-form podcasts, live video streams, and interactive town halls, context tools must operate natively in live audio visual environments.

From Browser Extensions to Native Ambient Interfaces

The Chrome extension is merely a stepping stone. As computing paradigms shift away from traditional screens, tools like InTruth will migrate directly into native hardware and ubiquitous spatial interfaces:

  • Spatial Computing & AR Glasses: In augmented reality environments, real-time verification layers will move from screen overlays to ambient heads-up displays during live, in-person events, political rallies, or public lectures.
  • Smart TV and Broadcast Integration: Streaming platforms and television hardware manufacturers will embed native “truth engines” directly into set-top boxes and smart OS environments, allowing viewers to toggle contextual overlays with a button on their remote.
  • Enterprise Video Conferencing: In business settings, real-time verification tools will be integrated into platforms like Zoom and Teams, serving as automated compliance and fact-verification assistants during corporate board meetings, earnings calls, and sales presentations.

The Strategic Counter-Maneuver: Adaptations in Public Rhetoric

Whenever a transformative technology alters the dynamic between speaker and audience, the speaker’s behavior adapts. The widespread adoption of live, automated truth engines will trigger a fundamental evolution in public rhetoric and media strategy:

  • Rhetorical Obfuscation vs. Precision: Bad actors will adapt by shifting from falsifiable statements to hyper-vague, highly emotive language designed to bypass database queries entirely. Conversely, transparent communicators will structure their speeches to explicitly cite primary sources in real time, encouraging instant automated validation.
  • Source Poisoning and Database Wars: The strategic battleground will shift from the speaker’s podium to the underlying reference databases. Public figures and interest groups will actively attempt to flood primary source repositories or manipulate public records to influence real-time algorithmic outputs.

Ultimately, InTruth is not just a tool for today’s web browsing — it is a prototype for the ambient, real-time context infrastructure that will define the future of human communication and public accountability.

Section V: The Verdict — Innovation or Novelty?

When we weigh InTruth against our core criteria for human-centered change, a clear distinction emerges: technological capability alone does not equal innovation. A product can feature cutting-edge machine learning and instantaneous audio transcription, yet remain an intrusive novelty if it fails to solve the broader human experience challenge.

To determine whether InTruth represents a category-defining breakthrough or just another temporary browser plugin, we must compare the characteristics of an incremental feature against those of a true, human-centered innovation:

Dimension Incremental Feature (Novelty) True Human-Centered Innovation
Core Focus Speed, raw transcription accuracy, and automated database lookups. Trust, context retention, and enhancing human comprehension without fatigue.
User Experience Intrusive pop-up overlays that interrupt viewing flow and increase cognitive friction. Non-intrusive micro-insights that empower individual judgment smoothly and ambiently.
Perceived Value An authoritative “referee” dictating rigid binary judgments (True/False). A transparent research assistant providing accessible primary-source receipts.
Systemic Impact A niche fact-checking widget for political hobbyists and journalists. A foundational layer for public accountability in the attention economy.

The Final Verdict

InTruth possesses all the technical ingredients necessary to become a transformative tool. However, its ultimate success will not be decided in the code base — it will be decided in the user experience layer.

If InTruth focuses merely on real-time execution speed and automated scorekeeping, it risks becoming a distraction that users disable after the initial novelty wears off. But if the team designs for low cognitive load, prioritizes source transparency over judgmental scoring, and seamlessly embeds context into the user’s natural viewing flow, InTruth will move from a clever browser extension to a vital engine of trust in the modern digital ecosystem.

Section VI: Strategic Recommendations for the InTruth Team

To successfully cross the chasm from a promising technical prototype to a category-defining human-centered innovation, the InTruth team must execute on four strategic imperatives. These recommendations focus on maximizing value translation, optimizing experience design, and building systemic trust across the entire ecosystem.

1. Prioritize Sourcing Over Scoring

Avoid the trap of playing “automated referee.” Assigning rigid, top-down truth scores or binary labels creates immediate psychological friction and invites accusations of algorithmic bias. Instead, shift the design emphasis entirely toward rapid, neutral receipt delivery:

  • Provide Direct Evidence: Present exact quotes from official public records, legislative documents, or peer-reviewed data alongside the spoken transcript.
  • Highlight Context Gaps: Explicitly show what information was omitted (e.g., “Stat represents 2021 data, excluding 2024 updates”) rather than declaring a statement flatly false.
  • Empower Human Judgment: Frame every overlay as an objective research receipt that respects the viewer’s intelligence to draw their own conclusion.

2. Empower the Nine Innovation Roles

Systemic adoption requires engaging users across distinct behavioral profiles. Product design should specifically accommodate the full spectrum of user roles from the Nine Innovation Roles, particularly:

  • The Troubleshooter: Provide deep-dive analytical modes allowing researchers, journalists, and policy analysts to inspect raw source metadata, API logs, and verification pipelines in real time.
  • The Customer Champion: Ensure the experience is continuously tuned to human needs, minimizing cognitive friction and preventing outrage fatigue so the tool feels like a trusted personal assistant.
  • The Evangelist: Build friction-free “shareable receipt” mechanics, enabling everyday users to export verified video clips and primary-source overlays directly to social networks with a single click.

3. Design for Zero-Friction Cognitive Ergonomics

The ultimate goal of experience design is making powerful capabilities feel effortless. If using InTruth requires conscious effort, adoption will stall among mainstream audiences:

  • Ambient Micro-Interactions: Keep overlays compact, glanceable, and peripheral to the primary video frame. Use non-intrusive status indicators that expand only upon intent.
  • Smart Thresholding: Implement confidence scoring models behind the scenes so the UI only interrupts the viewer when high-value, high-certainty discrepancies occur.
  • Zero-Setup Onboarding: Ensure the extension works immediately out of the box on major video platforms without requiring complex filter configurations or custom API setups.

4. Institutionalize Radical Transparency and Decentralized Trust

In an environment marked by deep institutional skepticism, the platform itself must be beyond reproach. Trust cannot be requested; it must be structurally demonstrated:

  • Open-Source Verification Logic: Make the underlying matching algorithms, prompt architectures, and scoring heuristics publicly auditable on GitHub.
  • Pluralistic Primary Databases: Draw from a diverse, transparently published set of non-partisan archives, government databases, and academic repositories rather than proprietary black-box datasets.
  • Community Audit Mechanisms: Allow trusted independent researchers and public ombudsmen to review, flag, and continuously refine source mappings.

Frequently Asked Questions

Is InTruth a fact-checker or an automated referee?

InTruth functions as a real-time research assistant rather than a judgmental referee. Instead of assigning arbitrary “True” or “False” ratings, it provides instant primary-source receipts — such as official government records, voting histories, and economic datasets — so users can verify live statements and draw their own conclusions without leaving their video stream.

How does InTruth reduce cognitive overload during live videos?

By shifting from active searching to ambient discovery, InTruth eliminates the friction of pausing videos or opening secondary browser tabs. Its glanceable, non-intrusive micro-overlays deliver contextual highlights in real time, preserving the viewer’s natural flow while ensuring critical facts are instantly accessible.

Why is experience design critical for automated fact-checking tools?

Technology enables capability, but experience design enables adoption. If a fact-checking tool uses intrusive pop-ups or accusatory language, it triggers user fatigue or cognitive defensiveness. Human-centered design ensures the tool delivers high value with zero setup effort, low cognitive friction, and maximum source transparency.


Image credits: Gemini

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Google Gemini to clean up the article, add images and create infographics.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Constrained Innovation is Beating Unconstrained Innovation – Again

Leave a reply

Constrained Innovation is Beating Unconstrained Innovation - Again

by Braden Kelley and Art Inteligencia

Every few years, Silicon Valley rediscovers a lesson the rest of the innovation world already knows: constraints don’t kill breakthroughs — they focus them.

This week’s AI headlines make the point again. Moonshot’s Kimi K3 and Thinking Machines Lab’s Inkling are not “unlimited compute with unlimited budget” stories. They are constrained-innovation stories — open-weight models built to compete with (and sometimes beat) far richer, closed frontier systems from OpenAI and Anthropic on the tasks that matter to builders. At the same time, a quieter race is packing surprising capability into models small enough to live on a smartphone with 6 GB of RAM or less.

If you lead change, product, or experience design, this is not just a model-release week. It is a reminder of how innovation actually works when resources are scarce, goals are clear, and “more” is not allowed to substitute for “better.”

The unconstrained myth

Unconstrained innovation sounds romantic: infinite GPUs, infinite capital, infinite permission to chase every benchmark.

In practice, unconstrained environments often produce:

  • Feature sprawl instead of sharp value
  • Capability inflation instead of usable outcomes
  • Vendor dependence instead of organizational learning
  • Status races (who has the biggest model) instead of customer impact

Constrained innovation does the opposite. It forces tradeoffs. Tradeoffs force clarity. Clarity forces design.

We’ve seen this movie before — in lean startups, in frugal engineering, in design-to-cost product development, in wartime R&D. The pattern is durable:

When you cannot buy your way to “more,” you must invent your way to “enough.”

AI is now teaching that lesson at planetary scale.

Kimi K3: open weight, frontier pressure

China’s Moonshot AI released Kimi K3 in mid-July 2026 as what it calls the world’s first open ~3T-class model — roughly 2.8 trillion parameters, native vision, and a 1-million-token context window, with full weights promised for public release.

Be precise about the scoreboard, because hype helps no one:

  • Moonshot itself says K3’s overall performance still trails Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6 Sol.
  • On multiple evaluations, though, K3 is competitive with — and on some coding, agent, long-horizon engineering, and frontend-building tasks ahead of — strong closed models sitting just behind the absolute tip of the spear.
  • Independent evaluators have placed it near GPT-5.5 / Claude Opus-class systems on several complex multi-step workloads, while still acknowledging Fable 5 as the tougher overall ceiling.

That combination is the real story: not “open models already own everything,” but “open models are close enough, open enough, and cheap enough to change the game.”

Constraint here is structural. Moonshot is not playing with the same geopolitical, capital, and closed-ecosystem advantages as the largest U.S. labs. So it optimized for:

  • Open weights (download, run, modify)
  • Architecture efficiency (MoE-style sparsity and novel attention choices)
  • Task-relevant dominance where developers actually feel pain (coding agents, long context, UI building)

That is constrained innovation: win where it matters for users, not where the press release wants a clean sweep.

Inkling: constraint as a product philosophy

Days earlier, Thinking Machines Lab — founded by former OpenAI CTO Mira Murati — released Inkling, its first open-weights model.

Inkling is a multimodal Mixture-of-Experts system (~975B total / ~41B active parameters), trained across text, images, audio, and video, with a large context window and Apache 2.0 weights on Hugging Face. Critically, the lab is not claiming Inkling is the strongest model available, open or closed.

Instead, Thinking Machines is making a different bet — one every human-centered innovator should recognize:

The winning model is not always the biggest generalist. It is the one an organization can shape.

Their framing is customization, efficient controllable “thinking effort,” and a base model designed to be adapted. Alongside Inkling they previewed Inkling-Small (lighter active-parameter footprint) for lower cost and latency.

This is constrained innovation as strategy:

  • Don’t outspend OpenAI/Anthropic on every frontier benchmark.
  • Out-enable customers on fit, control, and adaptation.
  • Treat “open weights + fine-tuning path” as the product, not a side quest.

In experience-design terms: they are optimizing for agency, not spectacle.

The pocket frontier: intelligence that fits in 6 GB

While the giants argue about trillion-parameter scoreboards, another constrained race is rewriting daily experience design: on-device AI.

Phones with ~6 GB of RAM are now practical homes for capable small language models — typically 1B–3B class models under aggressive 4-bit quantization, often with NPU acceleration (Apple Neural Engine, Qualcomm Hexagon, and peers). Families like Gemma’s efficient variants, Phi-class minis, Llama 3.2 small models, and Apple’s on-device foundation model path are not “tiny ChatGPT cosplay.” They are differently designed systems: distillation, quantization-aware training, sliding-window/grouped-query attention, and task specialization.

What becomes possible when intelligence must fit in a pocket?

  • Privacy by architecture (data never leaves the device)
  • Latency that feels like UI, not waiting for a cloud round trip
  • Offline resilience
  • Ambient assistance without a permanent surveillance subscription

This is FutureHacking in the literal sense: the future arriving first where constraint is non-negotiable — battery, thermal envelope, memory bandwidth, and user trust.

Unconstrained cloud models will still win the hardest reasoning contests for a while. Constrained on-device models will win moments — the thousands of tiny interactions that shape whether people feel helped or hunted by technology.

A simple framework: Three Arenas of Constrained AI Advantage

Leaders should stop asking only “Who has the best model?” and start asking which arena they are competing in:

  1. Frontier Arena — Absolute peak reasoning. Still often favors well-funded closed labs (Fable 5 / GPT-5.6 Sol class). Use sparingly for the hardest 10–20% of work.
  2. Open Adaptation Arena — Near-frontier capability + weights you can own, route, fine-tune, and host. Kimi K3 and Inkling are attacking this arena hard. Ideal for product teams, agents, and regulated environments.
  3. Edge Experience Arena — Models compressed into phone-scale memory. Wins on privacy, speed, cost-at-scale, and human experience continuity. This is where unconstrained cloud thinking often fails customers.

Constrained innovation beats unconstrained innovation when the arena rewards focus.

Implications for organizations (not just AI labs)

If you are charting change inside a company, the lesson is operational:

  1. Budget is a design tool. Cap tokens, latency, and model size early. Force product clarity.
  2. Route by job-to-be-done. Don’t send every prompt to the most expensive frontier model. Reserve it for true hard cases.
  3. Prefer adaptable over mythical “best.” An open model you can fine-tune to your workflow may outperform a slightly smarter generalist you can’t shape.
  4. Design for the edge. Anything frequent, personal, or privacy-sensitive should be a candidate for on-device or hybrid architectures.
  5. Measure outcomes, not vibes. Benchmarks matter; customer task completion, cost per successful outcome, and trust matter more.

This is human-centered change applied to AI portfolios: start from experience, not ego.

We’ve seen this movie — and the sequel is here

Constrained innovation beat unconstrained innovation in Japanese postwar manufacturing quality, Israeli “startup nation” necessity engineering, mobile-first product design in bandwidth-poor markets, and every great design brief that began with “You only get X.”

Now it is beating — or at least pressuring — unconstrained AI again.

Kimi K3 shows that open, resource-conscious frontier building can meet or beat closed leaders on key developer battlegrounds even while still trailing at the absolute peak. Inkling shows that refusing the one-size-fits-all arms race can itself be a strategy. Phone-scale models show that the most human future may be the one small enough to live beside us without phoning home.

The organizations that win the next decade will not be those with the least constraint.
They will be those who treat constraint as a creative operating system.

Because in innovation, as in life:

Limits don’t stop the future. They decide who gets there first — and who arrives with something people can actually use.

Image credits: Meta.AI

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Cursor to clean up the article.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

How to Calculate the ROI of Customer Experience

Leave a reply

Announcing the Launch of a Free CX ROI Calculator

by Braden Kelley

Most executive teams already believe that customer experience (CX) matters. Almost none of them can say, in dollars, what a specific improvement found in a Customer Experience Audit is worth — and that gap is usually the real reason a CX investment stalls before it reaches a budget conversation. Here’s a framework for closing it, backed by the research, plus a free calculator to run the numbers on your own business.

Why “CX matters” isn’t a business case

“Customer experience drives loyalty” is true, and it convinces almost no one holding a budget. What moves a budget conversation is a specific number: this metric, moved by this much, produces this many retained customers, worth this much revenue. Most CX teams never make that translation, so the initiative competes for funding against proposals that speak fluent finance while CX speaks fluent satisfaction score.

The good news is that the translation isn’t guesswork. There’s two decades of published research connecting experience metrics to financial outcomes — the trick is applying it to your own numbers instead of citing it as an abstract principle.

The research behind the number

Bain & Company, the originator of the Net Promoter Score, has found that NPS explains roughly 20% to 60% of the variation in organic growth rates between competitors in the same market, and that the NPS leader in a given industry typically outgrows competitors by more than double. In an early, widely cited analysis, Bain found that Dell’s detractors made up about 15% of its customer base and represented roughly $68 million in lost revenue — and estimated that converting just 2% to 8% of those detractors into promoters could add approximately $167 million in annual revenue.

7 pts → ~1%NPS increase → revenue growth (London School of Economics)
10 pts → 3.2%NPS increase → B2B upsell revenue (CustomerGauge)
5 pts → 25–95%Retention increase → profit increase (Reichheld / Bain)

A separate study from the London School of Economics found that a 7-point increase in NPS corresponds to roughly 1% revenue growth, while CustomerGauge’s research in B2B contexts found a tighter, more immediate link: a 10-point NPS increase correlating with a 3.2% increase in upsell revenue among existing accounts. Underneath all of it sits Fred Reichheld’s original retention research at Bain, which found that a 5-point improvement in customer retention increases profits by 25% to 95%, depending on the industry and business model.

The wide range in that last figure isn’t a weakness in the research — it’s the whole point. The financial return on a CX improvement depends on your margin structure, your customer lifetime value, and how much of the retained revenue is truly incremental. A generic industry number can’t answer that. Your own numbers can.

The four-step value chain

Here’s the framework that turns the research above into a number specific to your business:

  1. Experience Metric — the number you already track: NPS, CES, or CSAT.
  2. Behavioral Outcome — the specific customer behavior that metric predicts: renewing, referring, buying again, or churning.
  3. Financial Outcome — that behavior’s dollar value: retained revenue, reduced cost-to-serve, lower acquisition cost.
  4. The Intervention — what actually has to change to move Box 1 in the first place: a redesigned onboarding flow, a fixed billing process, a retrained support tier.

CX ROI 4 Box Framework

This is also the fastest way to diagnose why a past CX initiative didn’t show up in revenue: almost always, it moved Box 1 (the score) without a demonstrated effect on Box 2 (a specific behavior), so it was never going to reach Box 3. The fix isn’t more CX effort in general — it’s picking an intervention with a direct, traceable line to a named behavior, and measuring that behavior directly.

What “typical” looks like, by industry

Churn rates and cost-to-serve vary meaningfully by industry — and by methodology, which is worth naming honestly rather than smoothing over. Here are representative midpoints reconciled across several published benchmark studies:

Industry Typical annual churn Cost per service contact
SaaS / Software 5–14% $18–$35
Retail / eCommerce 20–37% $2.70–$12
Financial Services / Insurance 15–20% $15–$25
Healthcare 7–9% $50–$60
Telecom / Utilities 15–25% $20–$30
B2B Professional Services ~10–13% $30–$60

These are starting points for a company with no internal baseline yet — not universal constants. The strongest version of any business case replaces these with your own churn rate, revenue per customer, and service cost the moment that data exists.

CX ROI Calculator
Want to skip straight to your own number? Use the free CX ROI Calculator →
It runs this exact framework against your own customer count, revenue per customer, and churn rate, and gives you a business-case-ready total in under two minutes.

How to build the business case, step by step

1. Start with your own numbers

Current churn rate, average revenue per customer, and cost per service contact — pulled from finance or CS systems — will always be more persuasive than an industry benchmark. Use published figures only where internal data doesn’t exist yet.

2. Separate the well-established link from your company-specific estimate

The macro relationship between experience and growth (Bain, LSE, CustomerGauge) is well documented and easy to defend by name. The precise dollar impact for your company is always a modeled estimate — say so explicitly, and the business case gains credibility rather than losing it.

3. Model conservatively, then show the range

A single point estimate invites a single objection. A modeled range — conservative, moderate, optimistic — tends to survive scrutiny far better, because it demonstrates the thinking rather than just the output.

4. Tie the number to a specific intervention

Executives fund actions, not scores. Pair the projected financial impact with the specific initiative expected to produce it, rather than presenting the improvement as if it happens on its own.

Try it on your own numbers

The fastest way to see this framework in action is to run it against your own business. The CX ROI Calculator uses the same four-step chain described above — enter your customer count, revenue per customer, and current churn rate (or start from an industry benchmark), and it estimates the annual revenue and cost-to-serve impact of a defined experience improvement, along with a summary you can paste straight into a slide.

Get the CX ROI Benchmark Report — the full industry benchmark table with sources, the CX Value Chain framework, and answers to the five objections a CFO is most likely to raise. Enter your email and we’ll send it straight to your inbox.

If after exploring the ROI calculator you would like to explore unlocking revenue opportunities for your business with a Customer Experience Audit, contact me directly. I’m happy to have a no-obligation conversation about whether an audit makes sense for your current situation.

Image Credit: Gemini

Content Authenticity Statement: The topic area, key elements to focus on, etc. were decisions made by Braden Kelley, with a little help from Claude to clean up the article and Gemini to add images.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.