Tag Archives: AI

Causal AI

Moving Beyond Prediction to Purpose

LAST UPDATED: February 13, 2026 at 5:13 PM

Causal AI

GUEST POST from Art Inteligencia

For the last decade, the business world has been obsessed with predictive models. We have spent billions trying to answer the question, “What will happen next?” While these tools have helped us optimize supply chains, they often fail when the world changes. Why? Because prediction is based on correlation, and correlation is not causation. To truly innovate using Human-Centered Innovation™, we must move toward Causal AI.

Causal AI is the next frontier of FutureHacking™. Instead of merely identifying patterns, it seeks to understand the why. It maps the underlying “wiring” of a system to determine how changing one variable will influence another. This shift is vital because innovation isn’t about following a trend; it’s about making a deliberate intervention to create a better future.

“Data can tell you that two things are happening at once, but only Causal AI can tell you which one is the lever and which one is the result. Innovation is the art of pulling the right lever.”
— Braden Kelley

The End of the “Black Box” Strategy

One of the greatest barriers to institutional trust is the “Black Box” nature of traditional machine learning. Causal AI, by its very nature, is explainable. It provides a transparent map of cause and effect, allowing human leaders to maintain autonomy and act as the “gardener” tending to the seeds of technology.

Case Study 1: Personalized Medicine and Healthcare

A leading pharmaceutical institution recently moved beyond predictive patient modeling. By using Causal AI to simulate “What if” scenarios, they identified specific causal drivers for individual patients. This allowed for targeted interventions that actually changed outcomes rather than just predicting a decline. This is the difference between watching a storm and seeding the clouds.

Case Study 2: Retail Pricing and Elasticity

A global retail giant utilized Causal AI to solve why deep discounts led to long-term dips in brand loyalty. Causal models revealed that the discounts were causing a shift in quality perception in specific demographics. By understanding this link, the company pivoted to a human-centered value strategy that maintained price integrity while increasing engagement.

Leading the Causal Frontier

The landscape of Causal AI is rapidly maturing in 2026. causaLens remains a primary pioneer with their Causal AI operating system designed for enterprise decision intelligence. Microsoft Research continues to lead the open-source movement with its DoWhy and EconML libraries, which are now essential tools for data scientists globally. Meanwhile, startups like Geminos Software are revolutionizing industrial intelligence by blending causal reasoning with knowledge graphs to address the high failure rate of traditional models. Causaly is specifically transforming the life sciences sector by mapping over 500 million causal relationships in biomedical data to accelerate drug discovery.

“Causal AI doesn’t just predict the future — it teaches us how to change it.”
— Braden Kelley

From Correlation to Causation

Predictive models operate on correlations. They answer: “Given the patterns in historical data, what will likely happen next?” Causal models ask a deeper question: “If we change this variable, how will the outcome change?” This fundamental difference elevates causal AI from forecasting to strategic influence.

Causal AI leverages counterfactual reasoning — the ability to simulate alternative realities. It makes systems more explainable, robust to context shifts, and aligned with human intentions for impact.

Case Study 3: Healthcare — Reducing Hospital Readmissions

A large health system used predictive analytics to identify patients at high risk of readmission. While accurate, the system did not reveal which interventions would reduce that risk. Nurses and clinicians were left with uncertainty about how to act.

By implementing causal AI techniques, the health system could simulate different combinations of follow-up calls, personalized care plans, and care coordination efforts. The causal model showed which interventions would most reduce readmission likelihood. The organization then prioritized those interventions, achieving a measurable reduction in readmissions and better patient outcomes.

This example illustrates how causal AI moves health leaders from reactive alerts to proactive, evidence-based intervention planning.

Case Study 4: Public Policy — Effective Job Training Programs

A metropolitan region sought to improve employment outcomes through various workforce programs. Traditional analytics identified which neighborhoods had high unemployment, but offered little guidance on which programs would yield the best impact.

Causal AI empowered policymakers to model the effects of expanding job training, childcare support, transportation subsidies, and employer incentives. Rather than piloting each program with limited insight, the city prioritized interventions with the highest projected causal effect. Ultimately, unemployment declined more rapidly than in prior years.

This case demonstrates how causal reasoning can inform public decision-making, directing limited resources toward policies that truly move the needle.

Human-Centered Innovation and Causal AI

Causal AI complements human-centered innovation by prioritizing actionable insight over surface-level pattern recognition. It aligns analytics with stakeholder needs: transparency, explainability, and purpose-driven outcomes.

By embracing causal reasoning, leaders design systems that illuminate why problems occur and how to address them. Instead of deploying technology that automates decisions, causal AI enables decision-makers to retain judgment while accessing deeper insight. This synergy reinforces human agency and enhances trust in AI-driven processes.

Challenges and Ethical Guardrails

Despite its potential, causal AI has challenges. It requires domain expertise to define meaningful variables and valid causal structures. Data quality and context matter. Ethical considerations demand clarity about assumptions, transparency in limitations, and safeguards against misuse.

Causal AI is not a shortcut to certainty. It is a discipline grounded in rigorous reasoning. When applied thoughtfully, it empowers organizations to act with purpose rather than default to correlation-based intuition.

Conclusion: Lead with Causality

In a world of noise, Causal AI provides the signal. It respects human autonomy by providing the evidence needed for a human to make the final call. As you look to your next change management initiative, ask yourself: Are you just predicting the weather, or are you learning how to build a better shelter?

Strategic FAQ

How does Causal AI differ from traditional Machine Learning?

Traditional Machine Learning identifies correlations and patterns in historical data to predict future occurrences. Causal AI identifies the functional relationships between variables, allowing users to understand the impact of specific interventions.

Why is Causal AI better for human-centered innovation?

It provides explainability. Because it maps cause and effect, human leaders can see the logic behind a recommendation, ensuring technology remains a tool for human ingenuity.

Can Causal AI help with bureaucratic corrosion?

Yes. By exposing the “why” behind organizational outcomes, it helps leaders identify which processes (the wiring) are actually producing value and which ones are simply creating friction.

Disclaimer: This article speculates on the potential future applications of cutting-edge scientific research. While based on current scientific understanding, the practical realization of these concepts may vary in timeline and feasibility and are subject to ongoing research and development.

Image credits: Google Gemini

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

Why We Love to Hate Chatbots

Leave a reply

Why We Love to Hate Chatbots

GUEST POST from Shep Hyken

More and more, brands are starting to get the chatbot “thing” right. AI is improving, and customers are realizing that a chatbot can be a great first stop for getting quick answers or resolving questions. After all, if you have a question, don’t you want it answered now?

In a recent interview, I was asked, “What do you love about chatbots?” That was easy. Then came the follow-up question, “What do you hate about chatbots?” Also easy. The truth is, chatbots can deliver amazing experiences. They can also cause just as much frustration as a very long phone hold. With that in mind, here are five reasons to love (and hate) chatbots:

Why We Love Chatbots

24/7 Availability: Chatbots are always on. They don’t sleep. Customers can get help at any time, even during holidays.
Fast Response: Instant answers to simple questions, such as hours of operation, order status and basic troubleshooting, can be provided with efficiency and minimal friction.
Customer Service at Scale: Once you set up a chatbot, it can handle many customers at once. Customers won’t have to wait, and human agents can focus on more complicated issues and problems.
Multiple Language Capabilities: The latest chatbots are capable of speaking and typing in many different languages. Whether you need global support or just want to cater to different cultures in a local area, a chatbot has you covered.
Consistent Answers: When programmed properly, a chatbot delivers the same answers every time.

Chatbots Shep Hyken Cartoon

Why We Hate Chatbots

AI Can’t Do Everything, but Some Companies Think It Can: This is what frustrates customers the most. Some companies believe AI and chatbots can do it all. They can’t, and the result is frustrated customers who will eventually move on to the competition.
A Lack of Empathy: AI can do a lot, but it can’t express true emotions. For some customers, care, empathy and understanding are more important than efficiency.
Scripted Retorts Feel Robotic: Chatbots often follow strict guidelines. That’s actually a good thing, unless the answers provided feel overly scripted and generic.
Hard to Get to a Human: One of the biggest complaints about chatbots is, “I just want to talk to a person.” Smart companies make it easy for customers to leave AI and connect to a human.
There’s No Emotional Connection to a Chatbot: You’ll most likely never hear a customer say, “I love my chatbot.” A chatbot won’t win your heart. In customer service, sometimes how you make someone feel is more important than what you say.

Chatbots are powerful tools, but they are not a replacement for human connection. The best companies use AI to enhance support, not replace it. When chatbots handle the routine issues and agents handle the more complex and human moments, that’s when customer experience goes from efficient to … amazing.

Image credits: Unsplash

Sign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Win Your Way to an AI Job

Leave a reply

Anduril’s AI Grand Prix: Racing for the Future of Work

LAST UPDATED: January 28, 2026 at 2:27 PM

Anduril's AI Grand Prix: Racing for the Future of Work

GUEST POST from Art Inteligencia

The traditional job interview is an antiquated artifact, a relic of a bygone industrial era. It often measures conformity, articulateness, and cultural fit more than actual capability or innovative potential. As we navigate the complexities of AI, automation, and rapid technological shifts, organizations are beginning to realize that to find truly exceptional talent, they need to look beyond resumes and carefully crafted answers. This is where companies like Anduril are not just iterating but innovating the very hiring process itself.

Anduril, a defense technology company known for its focus on AI-driven systems, recently announced its AI Grand Prix — a drone racing contest where the ultimate prize isn’t just glory, but a job offer. This isn’t merely a marketing gimmick; it’s a profound statement about their belief in demonstrated skill over credentialism, and a powerful strategy for identifying talent that can truly push the boundaries of autonomous systems. It epitomizes the shift from abstract evaluation to purposeful, real-world application, emphasizing hands-on capability over theoretical knowledge.

“The future of hiring isn’t about asking people what they can do; it’s about giving them a challenge and watching them show you.”

— Braden Kelley

Why Challenge-Based Hiring is the New Frontier

This approach addresses several critical pain points in traditional hiring:

Uncovering Latent Talent: Many brilliant minds don’t fit the mold of elite university degrees or polished corporate careers. Challenge-based hiring can surface individuals with raw, untapped potential who might otherwise be overlooked.
Assessing Practical Skills: In fields like AI, robotics, and advanced engineering, theoretical knowledge is insufficient. The ability to problem-solve under pressure, adapt to dynamic environments, and debug complex systems is paramount.
Cultural Alignment Through Action: Observing how candidates collaborate, manage stress, and iterate on solutions in a competitive yet supportive environment reveals more about their true cultural fit than any behavioral interview.
Building a Diverse Pipeline: By opening up contests to a wider audience, companies can bypass traditional biases inherent in resume screening, leading to a more diverse and innovative workforce.

Beyond Anduril: Other Pioneers of Performance-Based Hiring

Anduril isn’t alone in recognizing the power of real-world challenges to identify top talent. Several other forward-thinking organizations have adopted similar, albeit varied, approaches:

Google’s Code Jam and Hash Code

For years, Google has leveraged competitive programming contests like Code Jam and Hash Code to scout for software engineering talent globally. These contests present participants with complex algorithmic problems that test their coding speed, efficiency, and problem-solving abilities. While not always directly leading to a job offer for every participant, top performers are often fast-tracked through the interview process. This allows Google to identify engineers who can perform under pressure and think creatively, rather than just those who can ace a whiteboard interview. It’s a prime example of turning abstract coding prowess into a tangible demonstration of value.

Kaggle Competitions for Data Scientists

Kaggle, now a Google subsidiary, revolutionized how data scientists prove their worth. Through its platform, companies post real-world data science problems—from predicting housing prices to identifying medical conditions from images—and offer prize money, and often, connections to jobs, to the teams that develop the best models. This creates a meritocracy where the quality of one’s predictive model speaks louder than any resume. Many leading data scientists have launched their careers or been recruited directly from their performance in Kaggle competitions. It transforms theoretical data knowledge into demonstrable insights that directly impact business outcomes.

The Human Element in the Machine Age

What makes these initiatives truly human-centered? It’s the recognition that while AI and automation are transforming tasks, the human capacity for ingenuity, adaptation, and critical thinking remains irreplaceable. These contests aren’t about finding people who can simply operate machines; they’re about finding individuals who can teach the machines, design the next generation of algorithms, and solve problems that don’t yet exist. They foster an environment of continuous learning and application, perfectly aligning with the “purposeful learning” philosophy.

The Anduril AI Grand Prix, much like Google’s and Kaggle’s initiatives, de-risks the hiring process by creating a performance crucible. It’s a pragmatic, meritocratic, and ultimately more effective way to build the teams that will define the next era of technological advancement. As leaders, our challenge is to move beyond conventional wisdom and embrace these innovative models, ensuring we’re not just ready for the future of work, but actively shaping it.

Anduril Fury

Frequently Asked Questions

What is challenge-based hiring?

Challenge-based hiring is a recruitment strategy where candidates demonstrate their skills and problem-solving abilities by completing a real-world task, project, or competition, rather than relying solely on resumes and interviews.

What are the benefits of this approach for companies?

Companies can uncover hidden talent, assess practical skills, observe cultural fit in action, and build a more diverse talent pipeline by focusing on demonstrable performance.

How does this approach benefit candidates?

Candidates get a fair chance to showcase their true abilities regardless of traditional credentials, gain valuable experience, and often get direct access to influential companies and potential job offers based purely on merit.

Image credits: Wikimedia Commons, Google Gemini

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

We Must Hold AI Accountable

Leave a reply

We Must Hold AI Accountable

GUEST POST from Greg Satell

About ten years ago, IBM invited me to talk with some key members on the Watson team, when the triumph of creating a machine that could beat the best human players at the game show Jeopardy! was still fresh. I wrote in Forbes at the time that we were entering a new era of cognitive collaboration between humans, computers and other humans.

One thing that struck me was how similar the moment seemed to how aviation legend Chuck Yeager described the advent of flying-by-wire, four decades earlier, in which pilots no longer would operate aircraft, but interface with a computer that flew the plane. Many of the macho “flyboys” weren’t able to trust the machines and couldn’t adapt.

Now, with the launch of ChatGPT, Bill Gates has announced that the age of AI has begun and, much like those old flyboys, we’re all going to struggle to adapt. Our success will not only rely on our ability to learn new skills and work in new ways, but the extent to which we are able to trust our machine collaborators. To reach its potential, AI will need to become accountable.

Recognizing Data Bias

With humans, we work diligently to construct safe and constructive learning environments. We design curriculums, carefully selecting materials, instructors and students to try and get the right mix of information and social dynamics. We go to all this trouble because we understand that the environment we create greatly influences the learning experience.

Machines also have a learning environment called a “corpus.” If, for example, you want to teach an algorithm to recognize cats, you expose it to thousands of pictures of cats. In time, it figures out how to tell the difference between, say, a cat and a dog. Much like with human beings, it is through learning from these experiences that algorithms become useful.

However, the process can go horribly awry. A famous case is Microsoft’s Tay, a Twitter bot that the company unleashed on the microblogging platform in 2016. In under a day, Tay went from being friendly and casual (“humans are super cool”) to downright scary, (“Hitler was right and I hate Jews”). It was profoundly disturbing.

Bias in the learning corpus is far more common than we often realize. Do an image search for the word “professional haircut” and you will get almost exclusively pictures of white men. Do the same for “unprofessional haircut” and you will see much more racial and gender diversity.

It’s not hard to figure out why this happens. Editors writing articles about haircuts portray white men in one way and other genders and races in another. When we query machines, we inevitably find our own biases baked in.

Accounting For Algorithmic Bias

A second major source of bias results from how decision-making models are designed. Consider the case of Sarah Wysocki, a fifth grade teacher who — despite being lauded by parents, students, and administrators alike — was fired from the D.C. school district because an algorithm judged her performance to be sub-par. Why? It’s not exactly clear, because the system was too complex to be understood by those who fired her.

Yet it’s not hard to imagine how it could happen. If a teacher’s ability is evaluated based on test scores, then other aspects of performance, such as taking on children with learning differences or emotional problems, would fail to register, or even unfairly penalize them. Good human managers recognize outliers, algorithms generally aren’t designed that way.

In other cases, models are constructed according to what data is easiest to acquire or the model is overfit to a specific set of cases and is then applied too broadly. In 2013, Google Flu Trends predicted almost double as many cases there actually were. What appears to have happened is that increased media coverage about Google Flu Trends led to more searches by people who weren’t sick. The algorithm was never designed to take itself into account.

The simple fact is that an algorithm must be designed in one way or another. Every possible contingency cannot be pursued. Choices have to be made and bias will inevitably creep in. Mistakes happen. The key is not to eliminate error, but to make our systems accountable through, explainability, auditability and transparency.

To Build An Era Of Cognitive Collaboration We First Need To Build Trust

In 2020, Ofqual, the authority that administers A-Level college entrance exams in the UK, found itself mired in scandal. Unable to hold live exams because of Covid-19, it designed and employed an algorithm that based scores partly on the historical performance of the schools students attended with the unintended consequence that already disadvantaged students found themselves further penalized by artificially deflated scores.

The outcry was immediate, but in a sense the Ofqual case is a happy story. Because the agency was transparent about how the algorithm was constructed, the source of the bias was quickly revealed, corrective action was taken in a timely manner, and much of the damage was likely mitigated. As Linus’s Law advises, “given enough eyeballs, all bugs are shallow.”

The age of artificial intelligence requires us to collaborate with machines, leveraging their capabilities to better serve other humans. To make that collaboration successful, however, it needs to take place in an atmosphere of trust. Machines, just like humans, need to be held accountable, their decisions and insights can’t be a “black box.” We need to be able to understand where their judgments come from and how they’re decisions are being made.

Senator Schumer worked on legislation to promote more transparency in 2024, but that is only a start and the new administration has pushed the pause button on AI regulation. The real change has to come from within ourselves and how we see our relationships with the machines we create. Marshall McLuhan wrote that media are extensions of man and the same can be said for technology. Our machines inherit our human weaknesses and frailties. We need to make allowances for that.

— Article courtesy of the Digital Tonto blog
— Image credit: Flickr

Sign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Solving the AI Trust Imperative with Provenance

Leave a reply

The Digital Fingerprint

LAST UPDATED: January 5, 2026 at 3:33 PM

The Digital Fingerprint - Solving the Trust Imperative with Provenance

GUEST POST from Art Inteligencia

We are currently living in the artificial future of 2026, a world where the distinction between human-authored and AI-generated content has become practically invisible to the naked eye. In this era of agentic AI and high-fidelity synthetic media, we have moved past the initial awe of creation and into a far more complex phase: the Trust Imperative. As my friend Braden Kelley has frequently shared in his keynotes, innovation is change with impact, but if the impact is an erosion of truth, we are not innovating — we are disintegrating.

The flood of AI-generated content has created a massive Corporate Antibody response within our social and economic systems. To survive, organizations must adopt Generative Watermarking and Provenance technologies. These aren’t just technical safeguards; they are the new infrastructure of reality. We are shifting from a culture of blind faith in what we see to a culture of verifiable origin.

“Transparency is the only antidote to the erosion of trust; we must build systems that don’t just generate, but testify. If an idea is a useful seed of invention, its origin must be its pedigree.” — Braden Kelley

Why Provenance is the Key to Human-Centered Innovation™

Human-Centered Innovation™ requires psychological safety. In 2026, psychological safety is under threat by “hallucinated” news, deepfake corporate communiques, and the potential for industrial-scale intellectual property theft. When people cannot trust the data in their dashboards or the video of their CEO, the organizational “nervous system” begins to shut down. This is the Efficiency Trap in its most dangerous form: we’ve optimized for speed of content production, but lost the efficiency of shared truth.

Provenance tech — specifically the C2PA (Coalition for Content Provenance and Authenticity) standards — allows us to attach a permanent, tamper-evident digital “ledger” to every piece of media. This tells us who created it, what AI tools were used to modify it, and when it was last verified. It restores the human to the center of the story by providing the context necessary for informed agency.

Case Study 1: Protecting the Frontline of Journalism

The Challenge: In early 2025, a global news agency faced a crisis when a series of high-fidelity deepfake videos depicting a political coup began circulating in a volatile region. Traditional fact-checking was too slow to stop the viral spread, leading to actual civil unrest.

The Innovation: The agency implemented a camera-to-cloud provenance system. Every image captured by their journalists was cryptographically signed at the moment of capture. Using a public verification tool, viewers could instantly see the “chain of custody” for every frame.

The Impact: By 2026, the agency saw a 50% increase in subscriber trust scores. More importantly, they effectively “immunized” their audience against deepfakes by making the absence of a provenance badge a clear signal of potential misinformation. They turned the Trust Imperative into a competitive advantage.

Case Study 2: Securing Enterprise IP in the Age of Co-Pilots

The Challenge: A Fortune 500 manufacturing firm found that its proprietary design schematics were being leaked through “Shadow AI” — employees using unauthorized generative tools to optimize parts. The company couldn’t tell which designs were protected “useful seeds of invention” and which were tainted by external AI data sets.

The Innovation: They deployed an internal Generative Watermarking system. Every output from authorized corporate AI agents was embedded with an invisible, robust watermark. This watermark tracked the specific human prompter, the model version, and the internal data sources used.

The Impact: The company successfully reclaimed its IP posture. By making the origin of every design verifiable, they reduced legal risk and empowered their engineers to use AI safely, fostering a culture of Human-AI Teaming rather than fear-based restriction.

Leading Companies and Startups to Watch

As we navigate 2026, the landscape of provenance is being defined by a few key players. Adobe remains a titan in this space with their Content Authenticity Initiative, which has successfully pushed the C2PA standard into the mainstream. Digimarc has emerged as a leader in “stealth” watermarking that survives compression and cropping. In the startup ecosystem, Steg.AI is doing revolutionary work with deep-learning-based watermarks that are invisible to the eye but indestructible to algorithms. Truepic is the one to watch for “controlled capture,” ensuring the veracity of photos from the moment the shutter clicks. Lastly, Microsoft and Google have integrated these “digital nutrition labels” across their enterprise suites, making provenance a default setting rather than an optional add-on.

Conclusion: The Architecture of Truth

To lead innovation in 2026, you must be more than a creator; you must be a verifier. We cannot allow the “useful seeds of invention” to be choked out by the weeds of synthetic deception. By embracing generative watermarking and provenance, we aren’t just protecting data; we are protecting the human connection that makes change with impact possible.

If you are looking for an innovation speaker to help your organization solve the Trust Imperative and navigate Human-Centered Innovation™, I suggest you look no further than Braden Kelley. The future belongs to those who can prove they are part of it.

Frequently Asked Questions

What is the difference between watermarking and provenance?

Watermarking is a technique to embed information (visible or invisible) directly into content to identify its source. Provenance is the broader history or “chain of custody” of a piece of media, often recorded in metadata or a ledger, showing every change made from creation to consumption.

Can AI-generated watermarks be removed?

While no system is 100% foolproof, modern watermarking from companies like Steg.AI or Digimarc is designed to be highly “robust,” meaning it survives editing, screenshots, and even re-recording. Provenance standards like C2PA use cryptography to ensure that if the data is tampered with, the “broken seal” is immediately apparent.

Why does Braden Kelley call trust a “competitive advantage”?

In a market flooded with low-quality or deceptive content, “Trust” becomes a premium. Organizations that can prove their content is authentic and their AI is transparent will attract higher-quality talent and more loyal customers, effectively bypassing the friction of skepticism that slows down their competitors.

Image credits: Google Gemini

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

Just Because You Can Use AI Doesn’t Mean You Should

1 Reply

Just Because You Can Use AI Doesn't Mean You Should

GUEST POST from Shep Hyken

I’m often asked, “What should AI be used for?” While there is much that AI can do to support businesses in general, it’s obvious that I’m being asked how it relates to customer service and customer experience (CX). The true meaning of the question is more about what tasks AI can do to support a customer, thereby potentially eliminating the need for a live agent who deals directly with customers.

First, as the title of this article implies, just because AI can do something, it doesn’t mean it should. Yes, AI can handle many customer support issues, but even if every customer were willing to accept that AI can deliver good support, there are some sensitive and complicated issues for which customers would prefer to talk to a human.

AI Shep Hyken Cartoon

Additionally, consider that, based on my annual customer experience research, 68% of customers (that’s almost seven out of 10) prefer the phone as their primary means of communication with a company or brand. However, another finding in the report is worth mentioning: 34% of customers stopped doing business with a company because self-service options were not provided. Some customers insist on the self-service option, but at the same time, they want to be transferred to a live agent when appropriate.

AI works well for simple issues, such as password resets, tracking orders, appointment scheduling and answering basic or frequently asked questions. Humans are better suited for handling complaints and issues that need empathy, complex problem-solving situations that require judgment calls and communicating bad news.

An AI-fueled chatbot can answer many questions, but when a medical patient contacts the doctor’s office about test results related to a serious issue, they will likely want to speak with a nurse or doctor, not a chatbot.

Consider These Questions Before Implementing AI For Customer Interactions

AI for addressing simple customer issues has become affordable for even the smallest businesses, and an increasing number of customers are willing to use AI-powered customer support for the right reasons. Consider these questions before implementing AI for customer interactions:

Is the customer’s question routine or fact-based?
Does it require empathy, emotion, understanding and/or judgment (emotional intelligence)?
Could the wrong answer cause a problem or frustrate the customer?
As you think about the reasons customers call, which ones would they feel comfortable having AI handle?
Do you have an easy, seamless way for the customer to be transferred to a human when needed?

The point is, regardless of how capable the technology is, it doesn’t mean it is best suited to deliver what the customer wants. Live agents can “read the customer” and know how to effectively communicate and empathize with them. AI can’t do that … yet. The key isn’t choosing between AI and humans. It’s knowing when to use each one.

Image credits: Google Gemini, Shep Hyken

Sign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Outcome-Driven Innovation in the Age of Agentic AI

Leave a reply

The North Star Shift

LAST UPDATED: January 5, 2026 at 5:29PM

Outcome-Driven Innovation in the Age of Agentic AI

by Braden Kelley

In a world of accelerating change, the rhetoric around Artificial Intelligence often centers on its incredible capacity for optimization. We hear about AI designing new materials, orchestrating complex logistics, and even writing entire software applications. This year, the technology has truly matured into agentic AI, capable of pursuing and achieving defined objectives with unprecedented autonomy. But as a specialist in Human-Centered Innovation™ (which pairs well with Outcome-Driven Innovation), I pose two crucial questions: Who is defining these outcomes, and what impact do they truly have on the human experience?

The real innovation of 2026 will show not just that AI can optimize against defined outcomes, but that we, as leaders, finally have the imperative — and the tools — to master Outcome-Driven Innovation and Outcome-Driven Change. If innovation is change with impact, then our impact is only as profound as the outcomes we choose to pursue. Without thoughtful, human-centered specifications, AI simply becomes the most efficient way to achieve the wrong goals, leading us directly into the Efficiency Trap. This is where organizations must overcome the Corporate Antibody response that resists fundamental shifts in how we measure success.

Revisiting and Applying Outcome-Driven Change in the Age of Agentic AI

As we integrate agentic AI into our organizations, the principles of Outcome-Driven Change (ODC) I first introduced in 2018 are more vital than ever. The core of the ODC framework rests on the alignment of three critical domains: Cognitive (Thinking), Affective (Feeling), and Conative (Doing). Today, AI agents are increasingly assuming the “conative” role, executing tasks and optimizing workflows at superhuman speeds. However, as I have always maintained, true success only arrives when what is being done is in harmony with what the people in the organization and customer base think and feel.

Outcome-Driven Change Framework

If an AI agent’s autonomous actions are misaligned with human psychological readiness or emotional context, it will trigger a Corporate Antibody response that kills innovation. To practice genuine Human-Centered Change™, we must ensure that AI agents are directed to pursue outcomes that are not just numerically efficient, but humanly resonant. When an AI’s “doing” matches the collective thinking and feeling of the workforce, we move beyond the Efficiency Trap and create lasting change with impact.

“In the age of agentic AI, the true scarcity is not computational power; it is the human wisdom to define the right ‘North Star’ outcomes. An AI optimizing for the wrong goal is a digital express train headed in the wrong direction – efficient, but ultimately destructive.” — Braden Kelley

From Feature-Building to Outcome-Harvesting

For decades, many organizations have been stuck in a cycle of “feature-building.” Product teams were rewarded for shipping more features, marketing for launching more campaigns, and R&D for creating more patents. The focus was on output, not ultimate impact. Outcome-Driven Innovation shifts this paradigm. It forces us to ask: What human or business value are we trying to create? What measurable change in behavior or well-being are we seeking?

Agentic AI, when properly directed, becomes an unparalleled accelerant for this shift. Instead of building a new feature and hoping it works, we can now tell an AI agent, “Achieve Outcome X for Persona Y, within Constraints Z,” and it will explore millions of pathways to get there. This frees human teams from the tactical churn and allows them to focus on the truly strategic work: deeply understanding customer needs, identifying ethical guardrails, and defining aspirational outcomes that genuinely drive Human-Centered Innovation™.

Case Study 1: Sustainable Manufacturing and the “Circular Economy” Outcome

The Challenge: A major electronics manufacturer in early 2025 aimed to reduce its carbon footprint but struggled with the complexity of optimizing its global supply chain, product design, and end-of-life recycling simultaneously. Traditional methods led to incremental, siloed improvements.

The Outcome-Driven Approach: They defined a bold outcome: “Achieve a 50% reduction in virgin material usage across all product lines by 2028, while maintaining profitability and product quality.” They then deployed an agentic AI system to explore new material combinations, reverse logistics networks, and redesign possibilities. This AI was explicitly optimized to achieve the circular economy outcome.

The Impact: The AI identified design changes that led to a 35% reduction in material waste within 18 months, far exceeding human predictions. It also found pathways to integrate recycled content into new products without compromising durability. The organization moved from a reactive “greenwashing” approach to proactive, systemic innovation driven by a clear, human-centric environmental outcome.

Case Study 2: Personalized Education and “Mastery Outcomes”

The Challenge: A national education system faced stagnating literacy rates, despite massive investments in new curricula. The focus was on “covering material” rather than ensuring true student understanding and application.

The Outcome-Driven Approach: They shifted their objective to “Ensure 90% of students achieve demonstrable mastery of core literacy skills by age 10.” An AI tutoring system was developed, designed to optimize for individual student mastery outcomes, rather than just quiz scores. The AI dynamically adapted learning paths, identified specific knowledge gaps, and even generated custom exercises based on each child’s learning style.

The Impact: Within two years, participating schools saw a 25% improvement in mastery rates. The AI became a powerful co-pilot for teachers, freeing them from repetitive grading and allowing them to focus on high-touch mentorship. This demonstrated how AI, directed by human-defined learning outcomes, can empower both educators and students, moving beyond the Efficiency Trap of standardized testing.

Leading Companies and Startups to Watch

As 2026 solidifies Outcome-Driven Innovation, several entities are paving the way. Amplitude and Pendo are evolving their product analytics to connect feature usage directly to customer outcomes. In the AI space, Anthropic‘s work on “Constitutional AI” is fascinating, as it seeks to embed human-defined ethical outcomes directly into the AI’s decision-making. Glean and Perplexity AI are creating agentic knowledge systems that help organizations define and track complex outcomes across their internal data. Startups like Metaculus are even democratizing the prediction of outcomes, allowing collective intelligence to forecast the impact of potential innovations, providing invaluable insights for human decision-makers. These players are all contributing to the core goal: helping humans define the right problems for AI to solve.

Conclusion: The Human Art of Defining the Future

The year 2026 is a pivotal moment. Agentic AI gives us unprecedented power to optimize, but with great power comes great responsibility — the responsibility to define truly meaningful outcomes. This is not a technical challenge; it is a human one. It requires deep empathy, strategic foresight, and the courage to challenge old metrics. It demands leaders who understand that the most impactful Human-Centered Innovation™ starts with a clear, ethically grounded North Star.

If you’re an innovation leader trying to navigate this future, remember: the future is not about what AI can do, but about what outcomes we, as humans, choose to pursue with it. Let’s make sure those outcomes serve humanity first.

Frequently Asked Questions

What is “Outcome-Driven Innovation”?

Outcome-Driven Innovation (ODI) is a strategic approach that focuses on defining and achieving specific, measurable human or business outcomes, rather than simply creating new features or products. AI then optimizes for these defined outcomes.

How does agentic AI change the role of human leaders in ODI?

Agentic AI frees human leaders from tactical execution and micro-management, allowing them to focus on the higher-level strategic work of identifying critical problems, understanding human needs, and defining the ethical, impactful outcomes for AI to pursue.

What is the “Efficiency Trap” in the context of AI and outcomes?

The Efficiency Trap occurs when AI is used to optimize for speed or cost without first ensuring that the underlying outcome is meaningful and human-centered. This can lead to highly efficient processes that achieve undesirable or even harmful results, ultimately undermining trust and innovation.

Image credits: Braden Kelley, Google 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.

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

Top 10 Human-Centered Change & Innovation Articles of December 2025

Leave a reply

Top 10 Human-Centered Change & Innovation Articles of December 2025 Drum roll please…

At the beginning of each month, we will profile the ten articles from the previous month that generated the most traffic to Human-Centered Change & Innovation. Did your favorite make the cut?

But enough delay, here are December’s ten most popular innovation posts:

Is OpenAI About to Go Bankrupt? — by Chateau G Pato
The Rise of Human-AI Teaming Platforms — by Art Inteligencia
11 Reasons Why Teams Struggle to Collaborate — by Stefan Lindegaard
How Knowledge Emerges — by Geoffrey Moore
Getting the Most Out of Quiet Employees in Meetings — by David Burkus
The Wood-Fired Automobile — by Art Inteligencia
Was Your AI Strategy Developed by the Underpants Gnomes? — by Robyn Bolton
Will our opinion still really be our own in an AI Future? — by Pete Foley
Three Reasons Change Efforts Fail — by Greg Satell
Do You Have the Courage to Speak Up Against Conformity? — by Mike Shipulski

BONUS – Here are five more strong articles published in November that continue to resonate with people:

Don’t Believe the Courageous Leadership Lie — by Robyn Bolton
Why 4D Printing is the Next Frontier of Human-Centered Change — by Art Inteligencia
Is There a Real Difference Between Leaders and Managers? — by David Burkus
Innovation Requires Defying Success — by Mike Shipulski
Leveraging the Power of Play to Innovate — by Janet Sernack

If you’re not familiar with Human-Centered Change & Innovation, we publish 4-7 new articles every week built around innovation and transformation insights from our roster of contributing authors and ad hoc submissions from community members. Get the articles right in your Facebook, Twitter or Linkedin feeds too!

Have something to contribute?

Human-Centered Change & Innovation is open to contributions from any and all innovation and transformation professionals out there (practitioners, professors, researchers, consultants, authors, etc.) who have valuable human-centered change and innovation insights to share with everyone for the greater good. If you’d like to contribute, please contact me.

P.S. Here are our Top 40 Innovation Bloggers lists from the last four years:

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

Can AI Replace the CEO?

Leave a reply

A Day in the Life of the Algorithmic Executive

LAST UPDATED: December 28, 2025 at 1:56 PM

Can AI Replace the CEO?

GUEST POST from Art Inteligencia

We are entering an era where the corporate antibody – that natural organizational resistance to disruptive change – is meeting its most formidable challenger yet: the AI CEO. For years, we have discussed the automation of the factory floor and the back office. But what happens when the “useful seeds of invention” are planted in the corner office?

The suggestion that an algorithm could lead a company often triggers an immediate emotional response. Critics argue that leadership requires soul, while proponents point to the staggering inefficiencies, biases, and ego-driven errors that plague human executives. As an advocate for Innovation = Change with Impact, I believe we must look beyond the novelty and analyze the strategic logic of algorithmic leadership.

“Leadership is not merely a collection of decisions; it is the orchestration of human energy toward a shared purpose. An AI can optimize the notes, but it cannot yet compose the symphony or inspire the orchestra to play with passion.”

— Braden Kelley

The Efficiency Play: Data Without Drama

The argument for an AI CEO rests on the pursuit of Truly Actionable Data. Humans are limited by cognitive load, sleep requirements, and emotional variance. An AI executive, by contrast, operates in Future Present mode — constantly processing global market shifts, supply chain micro-fluctuations, and internal sentiment analysis in real-time. It doesn’t have a “bad day,” and it doesn’t make decisions based on who it had lunch with.

Case Study 1: NetDragon Websoft and the “Tang Yu” Experiment

The Experiment: A Virtual CEO in a Gaming Giant

In 2022, NetDragon Websoft, a major Chinese gaming and mobile app company, appointed an AI-powered humanoid robot named Tang Yu as the Rotating CEO of its subsidiary. This wasn’t just a marketing stunt; it was a structural integration into the management flow.

The Results

Tang Yu was tasked with streamlining workflows, improving the quality of work tasks, and enhancing the speed of execution. Over the following year, the company reported that Tang Yu helped the subsidiary outperform the broader Hong Kong stock market. By serving as a real-time data hub, the AI signature was required for document approvals and risk assessments. It proved that in data-rich environments where speed of iteration is the primary competitive advantage, an algorithmic leader can significantly reduce operational friction.

Case Study 2: Dictador’s “Mika” and Brand Stewardship

The Challenge: The Face of Innovation

Dictador, a luxury rum producer, took the concept a step further by appointing Mika, a sophisticated female humanoid robot, as their CEO. Unlike Tang Yu, who worked mostly within internal systems, Mika serves as a public-facing brand steward and high-level decision-maker for their DAO (Decentralized Autonomous Organization) projects.

The Insight

Mika’s role highlights a different facet of leadership: Strategic Pattern Recognition. Mika analyzes consumer behavior and market trends to select artists for bottle designs and lead complex blockchain-based initiatives. While Mika lacks human empathy, the company uses her to demonstrate unbiased precision. However, it also exposes the human-AI gap: while Mika can optimize a product launch, she cannot yet navigate the nuanced political and emotional complexities of a global pandemic or a social crisis with the same grace as a seasoned human leader.

Leading Companies and Startups to Watch

The space is rapidly maturing beyond experimental robot figures. Quantive (with StrategyAI) is building the “operating system” for the modern CEO, connecting KPIs to real-work execution. Microsoft is positioning its Copilot ecosystem to act as a “Chief of Staff” to every executive, effectively automating the data-gathering and synthesis parts of the role. Watch startups like Tessl and Vapi, which are focusing on “Agentic AI” — systems that don’t just recommend decisions but have the autonomy to execute them across disparate platforms.

The Verdict: The Hybrid Future

Will AI replace the CEO? My answer is: not the great ones. AI will certainly replace the transactional CEO — the executive whose primary function is to crunch numbers, approve budgets, and monitor performance. These tasks are ripe for automation because they represent 19th-century management techniques.

However, the transformational CEO — the one who builds culture, navigates ethical gray areas, and creates a sense of belonging — will find that AI is their greatest ally. We must move from fearing replacement to mastering Human-AI Teaming. The CEOs of 2030 will be those who use AI to handle the complexity of the business so they can focus on the humanity of the organization.

Frequently Asked Questions

Can an AI legally serve as a CEO?

Currently, most corporate law jurisdictions require a natural person to serve as a director or officer for liability and accountability reasons. AI “CEOs” like Tang Yu or Mika often operate under the legal umbrella of a human board or chairman who retains ultimate responsibility.

What are the biggest risks of an AI CEO?

The primary risks include Algorithmic Bias (reinforcing historical prejudices found in the data), Lack of Crisis Adaptability (AI struggles with “Black Swan” events that have no historical precedent), and the Loss of Employee Trust if leadership feels cold and disconnected.

How should current CEOs prepare for AI leadership?

Leaders must focus on “Up-skilling for Empathy.” They should delegate data-heavy reporting to AI systems and re-invest that time into Culture Architecture and Change Management. The goal is to become an expert at Orchestrating Intelligence — both human and synthetic.

Image credits: Google Gemini

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