Brad Cleveland
Global Customer Experience and Service LeaderBeyond AI Optimism: What It Really Takes to Lead Customer Experience in 2026
“If you automate a broken process, you don’t fix it — you just get faster at going wrong.”
For over three decades, Brad Cleveland has watched the contact centre industry reinvent itself through every major technology wave — from the rise of CRM and the dot-com boom to mobile, analytics, and now generative AI.
But unlike previous cycles, Brad believes this moment is fundamentally different. Not because AI is replacing work overnight, but because it is now capable of participating directly in conversations, decisions, and workflows. That shift is forcing leaders to confront an uncomfortable reality: AI does not fix broken operations — it exposes them.
In this Expert Hub conversation, Brad unpacks why “AI optimism” can become dangerous when organisations ignore fundamentals, why many AI strategies still fail to start with customer needs, and why the future of CX will depend less on automation alone and more on clarity, governance, workforce design, and trust.
Q & A
"You became a founding partner of ICMI in 1991—you have watched every technology disruption cycle in contact centres for over three decades. You've seen the dot-com era, smartphones, big data, and now generative AI. Each wave brought predictions of mass job elimination that didn't materialise as forecast. What is genuinely different about this wave — and what mistake from previous cycles do you see leaders making again right now?"
What’s different is not just the technology, but the scope and speed. Generative AI can now participate directly in conversations, decisions, and workflows. That’s a step-change from prior waves that mainly enabled or routed work.
The biggest mistake I see leaders repeating is assuming technology drives outcomes on its own. We saw this in the dot-com era, with CRM, with mobile, and now again with AI. There’s a tendency to overestimate short-term impact and underestimate the importance of fundamentals.
The pattern is familiar: bold predictions of workforce reduction, followed by the reality that demand grows, complexity increases, and the work evolves rather than disappears.
What’s truly different this time is that AI amplifies whatever environment it’s placed into. In strong operations, it accelerates performance. In weak ones, it exposes and magnifies problems. So the risk isn’t just overestimating AI, it’s deploying it into systems that aren’t ready, and expecting it to compensate for that. That’s where leaders need to be more disciplined this time around.
SearchUnify Lens:
We agree with the view that AI amplifies the environment it is placed into; in strong operations, it accelerates performance, but in weak ones, it exposes and magnifies deep-seated problems. The primary antidote to this ‘weak operations’ risk is SearchUnify Federated Retrieval Augmented Generation (FRAG™).
Most AI initiatives fail because enterprise knowledge is siloed across various integrations such as Jira, Confluence, Slack, and Zendesk. Because AI is only as good as the data it can access, our platform unifies these silos into a single cognitive index. This ensures the AI isn’t just ‘generative’ – which often leads to making things up, but is instead ‘grounded’ in verified enterprise truth.
To ensure this grounding is both secure and reliable, we implement a robust Governance Layer. This layer addresses every stage of the AI lifecycle to prevent the common pitfalls of unmanaged AI deployment.
"You recently wrote that 'AI Optimism' may be compromising CX strategy, and the Wall Street Journal published your argument that companies most eager to use AI are often those that haven't fixed the fundamentals. What are the fundamentals you're referring to — and how does a leader know whether their organisation is ready for AI to add value, versus using AI to paper over problems that will eventually resurface?"
The fundamentals are surprisingly consistent across organizations: clear service strategy, well-defined processes, accurate knowledge, effective workforce management, and strong quality standards.
At a deeper level, it’s about clarity. Do you know what customers are trying to accomplish? Are your processes designed around that? Are your teams equipped to handle variability and complexity? A simple test I use is this: if you automate a broken process, do things get better, or just faster at going wrong?
Organizations ready for AI tend to have a few things in place. They understand demand and what’s driving interactions and why. They have reliable data and knowledge sources. And they have leadership alignment on what success looks like beyond cost reduction. Those that aren’t ready often use AI to “paper over” friction, hoping automation will reduce volume without addressing root causes. That rarely holds. The issues resurface, often at greater scale.
AI works best as an accelerator of clarity and discipline. It’s a terrible substitute for it.
SearchUnify Lens:
SearchUnify Lens: We agree that AI is an accelerator, not a cure; therefore, we combat the “papering over” risk by using Content Health and Auditing tools to ensure your data foundation is robust enough to support a successful transition from case-first to knowledge-first support.
"You coined the concept of 'fixed-pie thinking'— the tendency of organisations to inventory what they already do and look for ways to automate it, rather than starting with what customers actually need. How widespread is this mistake in your consulting work today, and what does a 'customer-needs-first' AI strategy actually look like when you're helping a leadership team build One?"
Fixed-pie thinking is extremely common. Most organizations start with what they already do and ask, “How can we automate this?” It’s an inside-out approach. A customer-needs-first strategy flips that. It starts with, what are customers trying to accomplish? Where is the effort high? Where does confusion or friction exist?
From there, you redesign the experience. Only then can you best determine where AI can help. In practice, this often leads to very different decisions. Instead of just deflecting contacts, you might eliminate them by fixing upstream issues. Instead of automating entire interactions, you might use AI to assist agents in high-complexity moments.
Here’s an example. Rather than building a chatbot to handle billing inquiries, a client redesigned the billing experience and reduced those contacts significantly. They then used AI to support the remaining complex cases. The shift is subtle but powerful. AI becomes part of a broader experience strategy, not the strategy itself.
SearchUnify Lens:
To move from “automation-first” to “needs-first,” we utilize LLM-powered Intent Detection to map the actual journey of a customer’s struggle, moving beyond simple keyword matching to true semantic understanding. This allows leadership to redesign the experience from the outside-in, deploying proactive support through Agentic AI that addresses the root cause of friction before it ever reaches the contact center, effectively shrinking the “pie” of avoidable effort.
"Your Customer Access Strategy is a ten-component blueprint that you've described as the guide for navigating AI decisions. Most organisations don't have anything like this — they're making AI investments ad hoc, channel by channel. What happens strategically when AI is deployed without this kind of framework underneath it, and what are the first questions a leader should be asking before committing to any AI investment?"
When AI is deployed without a unifying framework, organizations tend to optimize in silos. One team improves chat. Another invests in voice AI. Another adds automation to email. Each initiative may show progress, but the overall experience often becomes more fragmented.
From a customer’s perspective, it feels inconsistent. From an operational perspective, it creates duplication, gaps, and unintended consequences. You’ll often be shifting demand from one channel to another rather than reducing it.
A Customer Access Strategy brings coherence. It aligns channels, clarifies intent, and defines how work should flow across functional areas. Before committing to any AI investment, I encourage leaders to step back and ask a few fundamental questions:
- What are customers actually trying to accomplish and where is effort highest?
- What is driving demand, and how much of it is avoidable?
- How should work be distributed across channels, automation, and human support?
- What does “great” look like for this interaction, e.g., speed, accuracy, reassurance, resolution?
AI decisions make much more sense when they’re anchored in those answers. Otherwise, you risk building a faster, more complex version of what you already have.
SearchUnify Lens:
We solve the fragmentation Brad describes by using our FRAG™ framework to serve as the connective tissue for a Customer Access Strategy, ensuring that the same verified “enterprise brain” powers every channel. This architectural consistency prevents the common pitfall of “channel-specific silos,” providing a single cognitive search experience that maintains context as a customer moves from self-service to a live agent, eliminating the frustration of repeating information.
"You argue that the contact centre creates value on three levels — efficiency, customer satisfaction and loyalty, and strategic value through intelligence. In your experience, which of these three does AI investment most commonly target — and which does it most consistently fail to unlock? What is the structural barrier that stops organisations from capturing that third level of value?"
Most AI investment today is still aimed at efficiency, such as reducing handle time, deflecting contacts, lowering cost per interaction. That’s understandable. It’s measurable, immediate, and easy to justify.
Where organizations consistently fall short is the third level of value: strategic insight. The contact center is the richest source of customer intelligence in most organizations, but that value is often underleveraged. The structural barrier is that many organizations still view the contact center as an operational function, not a strategic one. Data is collected, but not translated into action across the business.
AI, especially interaction analytics, has the potential to change that by identifying patterns, root causes, and opportunities at scale. But it requires leadership alignment and processes to act on those insights. The biggest opportunity isn’t just doing the work faster. It’s using what you learn from the work to improve the business.
SearchUnify Lens:
The structural barrier to strategic intelligence is the inability to process unstructured data at scale. SearchUnify overcomes this through advanced search analytics that surface “Discovery Insights” – identifying unmet customer needs and emerging product friction points. By transforming raw interaction data into actionable tacit knowledge, we enable the contact center to provide “Level 3” value by delivering real-time feedback loops to Product, Engineering, and Marketing teams.
"You distinguish between 'foundation' quality standards — objective and measurable — and 'finesse' standards — the emotional and relational layer. As AI takes over more interactions, how should quality frameworks evolve? Can an AI interaction ever meet a finesse standard, or does finesse by definition require a human — and what does that mean for how organisations should be designing their hybrid human-AI service models?"
Foundation standards (encompassing accuracy, compliance, and correct processes) are well suited to AI. In many cases, AI can outperform humans in consistency.
Finesse is different. It involves judgment, empathy, tone, and adapting to nuance. AI is improving here, but it’s not the same as human connection, especially in complex or emotionally charged situations.
What this means is that quality frameworks need to become more explicit. Define what must always be right (foundation), and where flexibility and human judgment matter (finesse).
In hybrid models, AI can handle or support foundation elements, such as ensuring accuracy, surfacing knowledge, guiding next steps, while humans focus more on finesse. The goal isn’t to replicate humans with AI, but to combine strengths.
SearchUnify Lens:
We enable agents to focus on “finesse” by deploying SearchUnify Agent Helper to handle the high-volume cognitive load of foundation tasks—such as technical lookups, case summarization, and compliance checks. This optimizes the handoff between AI and humans, ensuring that while the AI handles the measurable “correctness,” the agent is free to provide the empathy and nuanced judgment that creates true customer loyalty.
"W. Edwards Deming told you early in your career to 'drive out the fear. ' You've carried that principle through decades of leadership consulting. In an AI-first environment, where agents fear replacement, leaders fear getting the strategy wrong, and customers fear being trapped in automated dead ends — what does 'driving out the fear' look like in practice for a contact centre leader in 2026?"
This principle is more relevant than ever. While Dr. Deming was primarily focused on fear among employees, it really involves employees, leaders, and customers. Each needs to be addressed directly.
For employees, fear often centers on replacement or lack of agency. The most effective leaders counter this with transparency and involvement. For example, they show where AI helps, where human skills matter more than ever, and they invest in developing those skills. When people see a future for themselves, and when they see the positive impact of their work, they engage differently.
For leaders, the fear is getting it wrong, especially investing time, resources and money heavily and missing the mark. The answer here is not to wait, but to take a disciplined approach: test, learn, iterate, and stay grounded in customer needs rather than hype.
For customers, a common fear is being trapped, unable to reach a human when it matters. This is critical. Organizations need to design clear, reliable paths to human support, especially for complex or high-stakes situations.
“Driving out fear” in this environment means replacing uncertainty with clarity. Clear expectations. Clear roles for AI and humans. Clear paths for customers. When that happens, trust builds, which ultimately determines whether AI succeeds or fails.
SearchUnify Lens:
Driving out fear requires transparency and agent agency, which is why we champion human-in-the-loop AI governance models that allow frontline staff to audit, refine, and “train” AI outputs in real-time. By giving agents the tools to oversee the AI, they transition from fearing replacement by the technology to becoming the master users who drive its evolution, fostering a culture of trust and technical competence.
"Your data shows that despite years of AI adoption, most contact centre leaders expect workloads to increase — and Gen Z is calling more than millennials. What does that tell us about the gap between what organisations expect AI to do and what customers actually want? And what does it mean for the organisations that have been cutting headcount based on deflection promises that haven't materialised?"
This is one of the most important signals in the market. Despite years of AI adoption, many leaders report increasing workloads. That tells us expectations and reality are misaligned. AI doesn’t just deflect demand, it often reveals and generates it. As access improves, customers engage more. As simple tasks are automated, remaining work becomes more complex.
We also see expanding channels, more complex products, and higher expectations for personalization. Gen Z, for example, will absolutely use digital channels, but they’ll also call when something matters.
For organizations that reduced headcount based on deflection assumptions, this creates pressure in the form of longer wait times, stressed teams, and inconsistent service. The takeaway is that AI should be planned as part of a broader demand and capacity strategy, not a shortcut to reducing resources. In many cases, the role of the agent becomes more important, not less.
SearchUnify Lens:
The paradox of “more AI = more work” is often due to Self-Service Deflection Failure. If a bot can’t solve a complex issue, it creates a frustrated customer who then calls in. SearchUnify’s FRAG™ technology ensures that self-service handles high-complexity tasks by grounding answers in technical documentation, not just FAQ fluff. This ensures that when a Gen Z customer uses digital channels, they actually get a resolution, not just a redirection.
More on: Why Traditional Deflection is Failing (And How to Fix It)
"Your newest LinkedIn Learning course is built specifically for frontline agents on how to work alongside AI. You've also just written about the five musts for building employee engagement in contact centres. How do you connect those two things — because most organisations treat AI adoption and employee engagement as separate workstreams, when in reality one directly determines the outcome of the other?"
These two are tightly linked, whether organizations recognize it or not. AI changes the nature of the work. It can reduce friction, by surfacing answers, automating after-call work, and providing real-time guidance. But it can also increase pressure if poorly implemented.
Engagement improves when agents feel equipped, trusted, and part of the process. That means involving them early, training them to work effectively with AI, and showing how it helps them succeed. In my work, the best-performing organizations treat AI as a tool for elevating the role. They invest in skills like judgment, communication, and problem-solving, alongside technical capability.
If AI is introduced as a cost-cutting tool, engagement suffers. If it’s introduced as a capability-building tool, engagement and performance improve. Ultimately, the experience you deliver to customers is directly tied to the experience you create for your people.
SearchUnify Lens:
We view AI as an “Employee Experience” (EX) tool first. A frustrated agent cannot deliver a great CX. By automating the “drudgery” – the searching, the tagging, the summarizing – we allow agents to engage in the meaningful work they were hired for. Our Escalation Prediction models also help managers intervene before an agent burns out, directly linking AI capability to employee retention.
Guide: The Link Between AI, Agent Experience, and Customer Loyalty
Looking Ahead
AI doesn't fix broken operations - it exposes them. The organizations that will lead in 2026 aren't those that moved fastest on AI adoption, but those that moved most deliberately. They asked hard questions first. They grounded AI in customer needs, not cost reduction. They treated workforce transition as a strategic priority. The leaders who get this right will view AI not as a technology shift, but as a discipline shift—in strategy, execution, and maintaining human values while scaling efficiency. The contact centre of 2026 won't look radically different from today. But the culture inside it will be transformed, not because the technology changed, but because leaders chose to use it as a tool for clarity, not a substitute for it.