voicebots | Human-Centered Change and Innovation

The Empathy Engine

LAST UPDATED: November 9, 2025 at 1:36PM
Re-engineering Trust and Retention in the AI Contact Center

by Braden Kelley

The contact center remains the single most critical point of human truth for a brand. It is where marketing promises meet operational reality. The challenge today, as highlighted by leaders like Bruce Gilbert of Young Energy at Customer Contact Week(CCW) in Nashville recently, is profound: Customers expect friction-less experiences with empathetic responses. The solution is not merely throwing technology at the problem; it’s about strategically weaving automation into the existing human fabric to create an Empathy Engine.

The strategic error most organizations make is starting with the technology’s capability rather than the human need. The conversation must start with empathy not the technology — focusing first on the customer and agent pain points. AI is not a replacement for human connection; it is an amplification tool designed to remove friction, build trust, and elevate the human agent’s role to that of a high-value relationship manager.

The Trust Imperative: The Cautious Adoption Framework

The first goal when introducing AI into the customer journey is simple: Building trust. The consumer public, after years of frustrating Interactive Voice Response (IVR) systems and rigid chatbots, remains deeply skeptical of automation. A grand, “all-in” AI deployment is often met with immediate resistance, which can manifest as call abandonment or increased churn.

To overcome this, innovation must adhere to a principle of cautious, human-centered rollout — a Cautious Adoption Framework: Starting small and starting with simple things can help to build this trust. Implement AI where the risk of failure is low and the utility is high — such as automating password resets, updating billing addresses, or providing initial diagnostics. These are the repetitive, low-value tasks that bore agents and frustrate customers. By successfully automating these simple, transactional elements, you build confidence in the system, preparing both customers and agents for more complex, AI-assisted interactions down the line. This approach honors the customer’s pace of change.

The Agent Retention Strategy: Alleviating Cognitive Load

The operational cost of the contact center is inextricably linked to agent retention. Finding and keeping high-quality agents remains a persistent challenge, primarily because the job is often highly stressful and repetitive. AI provides a powerful retention tool by directly addressing the root cause: cognitive load.

Reducing the cognitive load and stress level on agents is a non-negotiable step for long-term operational health. AI co-pilots must be designed to act as true partners, not simply data overlays. They should instantly surface relevant knowledge base articles, summarize the customer’s entire history before the agent picks up the call, or even handle real-time data entry. This frees the human agent to focus entirely on the empathetic response — active listening, problem-solving, and de-escalation. By transforming the agent’s role from a low-paid data processor into a high-value relationship manager, we elevate the profession, directly improving agent retention and turning contact center employment into an aspirational career path.

The Systemic Challenge: Orchestrating the AI Ecosystem

A major limiting factor in today’s contact center is the presence of fragmented AI deployments. Many organizations deploy AI in isolated pockets — a siloed chatbot here, a transcription service there. The future demands that we move far beyond siloed AI. The goal is complete AI orchestration across the enterprise, requiring us to get the AIs to talk to each other.

A friction-less customer experience requires intelligence continuity: a Voice AI must seamlessly hand off its collected context to a Predictive AI (which assesses the call risk), which then informs the Generative AI (that drafts the agent’s suggested response). This is the necessary chain of intelligence that supports friction-less service. Furthermore, complexity demands a blended AI approach, recognizing that the solution may involve more than one method (generative vs. directed).

For high-compliance tasks, a directed approach ensures precision: for instance, a flow can insert “read as is” instructions for regulatory disclosures, ensuring legal text is delivered exactly as designed. For complex, personalized problem-solving, a generative approach is vital. The best systems understand the regulatory and emotional context, knowing when to switch modes instantly and without customer intervention.

The Strategic Pivot: Investing in Predictive Empathy

The ultimate strategic advantage lies not in reacting to calls, but in preventing them. This requires a deeper investment in data science, moving from descriptive reporting on what happened to predictive analytics to understand why our customers are calling in before they dial the number.

This approach, which I call Predictive Empathy, uses machine learning to identify customers whose usage patterns, payment history, or recent service interactions suggest a high probability of confusion or frustration (e.g., first-time promotions expiring, unusual service interruptions). The organization then proactively initiates a personalized, AI-assisted outreach to address the problem or explain the confusion before the customer reaches the point of anxiety and makes the call. This shifts the interaction from reactive conflict to proactive support, immediately lowering call volume and transforming brand perception.

The Organizational Checkpoint: Post-Deployment Evolution

Once you’ve successfully implemented AI to address pain points, the work is not finished. A crucial strategic question must be addressed: What happens after AI deployment? What’s your plan?

As AI absorbs simple transactions, the nature of the calls that reach the human agent becomes disproportionately more complex, emotional, and high-value. This creates a skills gap in the remaining human workforce. The organization must plan for and fund the Up-skilling Initiative necessary to handle these elevated interactions, focusing on conflict resolution, complex sales, and deep relationship management. The entire organizational structure — training programs, compensation models, and career paths — must evolve to support this higher-skilled human workforce. By raising the value of the human role, the contact center transitions from a cost center into a profit-generating Relationship Hub.

Conclusion: Architecting the Human Layer

The goal of innovation in the contact center is not the elimination of the human, but the elevation of the human. By using AI to build trust, reduce cognitive load, enable predictive empathy, and connect disparate systems, we free the human agent to deliver on the fundamental customer expectation: a friction-less experience coupled with an empathetic response. This is how we re-engineer the contact center from a cost center into a powerful engine for talent retention and customer loyalty.

“AI handles the transaction. The human handles the trust. Design your systems to protect both.” — Braden Kelley

Your first step into the Empathy Engine: Map the single most stressful task for your top-performing agent and commit to automating 80% of its cognitive load using a simple AI co-pilot within the next 90 days.

What is that task for your organization?

Image credits: Google Gemini

Content Authenticity Statement: The topic area, key elements to focus on, insights captured from the Customer Contact Week session, panelists to mention, 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.

I’m Sorry Dave, But I Can’t Do That

LAST UPDATED: November 3, 2025 at 4:50PM

Why Going AI Only is Dumb

by Braden Kelley

Last month I had the opportunity to attend Customer Contact Week (CCW) in Nashville, Tennessee and following up on my article The Voicebots Are Coming I’d like to dig into the idea that companies like Klarna explored of eliminating all humans from contact centers. After all, what could possibly go wrong?

When I first heard that Klarna was going to eliminate humans from their contact centers and go all in on artificial intelligence I thought to myself that they would likely live to regret it. Don’t get me wrong, artificial intelligence (AI) voicebots and chatbots can be incredibly useful, and that proves out in the real world according to conference speakers that almost half of Fanatics calls are automated on the phone without getting to an agent. A lot of people are experimenting with AI but AI is no longer experimental. What Klarna learned is that when you choose to use AI to reduce your number of human agents, then if the AI is down you don’t have the ability anymore to just call in off duty agents to serve your customers.

But, on the flip side we know that having AI customer service agents as part of your agent mix can have very positive impacts on the business. Small businesses like Brothers That Just Do Gutters have found that using AI agents increased their scheduling of estimate visits over humans alone. National Debt Relief automated their customer insufficient funds (CIF) calls and added an escalation path (AI then agent) that delivered a 20% revenue lift over their best agents. They found that when an agent gets a NO, there isn’t much of an escalation path left. And, the delicate reality is that some people feel self conscious calling a human to talk about debt problems, and there may be other sensitive issues where callers would actually feel more comfortable talking to a voicebot than a human. In addition, Fanatics is finding that AI agents are resolving some issues FASTER than human agents. Taken together these examples show that often a hybrid approach (humans plus AI) yields better results than humans only or AI only, so design your approach consciously.

Now let’s look at some important statistics from Customer Management Practice research:

2/3 of people prefer calling in and talking by phone, but most of that is 55+ and the preference percentage declines every ten years younger you go until 30% for 18-24
3/4 of executives say more people want self service now than three years ago
3/4 of people want to spend less time getting support – so they can get back to the fun stuff, or back to business

Taken together these statistics help make the case for increasing the use of AI agents in the contact center. If you happen to be looking to use AI agents in servicing your customers (or even if you already are) then it is important to think about how you can use them to remove friction from the system and to strategically allocate your humans towards things that only humans can do. And if you need to win support from someone to go big with AI voicebots then pick an important use case instead of one that nobody cares about OR even better, pick something that you couldn’t have done before (example: a ride sharing company had AI voicebots make 5 million calls to have drivers validate their tax information).

Finally, as I was listening to some of these sessions it reminded me of a time when I was tasked with finding a new approach for staffing peak season for one of the Blue Cross/Blue Shield companies in the United States. At that time AI voicebots weren’t a thing and so I was looking at how we could partner with a vendor to have a small number of their staff on hand throughout the year and then rely on them to staff and train seasonal staff using those seasoned vendor staff instead of taking the best employees off the phone to train temps.

Even now, all contact centers will still need a certain level of human staffing. But, AI voicebots, AI simulation training for agents, and other new AI powered tools represent a great opportunity for creating a better solution for peak staffing in a whole host of industries with very cyclical contact demand that is hard to staff for. One example of this from Customer Contact Week was a story about how Fanatics must 5x their number of agents during high seasons and in practice this often results in their worst agents (temps they hired only for the season) serving some of their best customers (high $$ value clients).

Conclusion

AI voicebots can be a great help during demand peaks and other AI powered tools (QA, simulations, coaching, etc.) can help accelerate and optimize both your on-boarding of full-time agents, but also of seasonal agents as well. But don’t pare back your human agent pool too far!

What has been your experience with balancing human and AI agents?