Contact centers across financial services are quietly becoming the industry’s most expensive governance liability. The Consumer Financial Protection Bureau issued more than $3.1 billion in consumer relief orders in 2024 alone, and a significant portion of the underlying complaints trace back to unresolved service journeys — customers who contacted an institution repeatedly, escalated, and eventually filed. The leaders running these operations are not unaware of the problem. Most are simply solving it in the wrong sequence.
That, at least, is the diagnosis of David Fapohunda, a contact center operations executive with decades of large-scale transformation experience across financial services. His reframe is deceptively simple. “I stopped asking: ‘How do we reduce cost without hurting experience?’” Fapohunda explains, “and started asking: ‘What does this customer’s journey actually cost us?’ and ‘Where is the value leaking?’ Those two questions change everything about how you structure the work, what you measure, and where you invest.”
The Technology-First Model Is the Industry’s Most Expensive Assumption
The dominant pattern in contact center transformation is technology-first. New platforms, AI deployments, and contact center as a service (CCaaS) migrations get layered onto legacy organizational structures. Teams remain organized by channel, queue, or product silo. The expected returns fail to materialize. Industry data shows roughly 80% of contact centers are accelerating digital transformation. A much smaller share is seeing the returns. According to Fapohunda, the constraint is almost never the technology. It is the operating model underneath it.
His approach inverts the conventional sequence. Before making a single technology investment, Fapohunda restructures operations around customer journeys. The distinction matters more than it might initially appear. Channel-organized teams optimize for their channel. Journey-organized teams optimize for the underlying customer intent. When a single team owns the end-to-end journey — the escalations, the repeat contacts, the abandonment — those become their problem to solve rather than someone else’s to deflect.
Case data from local government operations that restructured around journeys in a single department delivered up to 40% first contact resolution improvement before any major technology investment — a pattern Fapohunda says he has seen replicated consistently across sectors. “I identify the three to five highest-friction journeys by repeat contact rate and effort,” he explains. “I stand up cross-functional pods around them. I give each pod end-to-end accountability and a balanced scorecard. Then, and only then, I layer in the technology investments that scale what is working.”
Repeat Contact Rate Cannot Be Gamed — Which Is Exactly Why It Should Lead Your Scorecard
The measurement model most operations are running, Fapohunda argues, is actively working against them. Tracking sixteen KPIs with roughly equal weight means prioritizing nothing. Average handle time and first contact resolution have been gamed for years. His preferred scorecard is tighter and more honest: FCR, CSAT, customer effort score, repeat contact rate, and cost per resolution — with explicit acknowledgment of the tradeoffs between them.
Repeat contact rate is the metric he trusts most. “Customers calling back about the same issue cannot be gamed,” he states. “It is the most honest signal in the operation about whether the underlying problem is actually being solved or merely handled.” He pairs it with customer effort score rather than CSAT as his leading indicator for loyalty in transactional environments, on the basis that effort predicts future behavior in ways satisfaction scores frequently do not.
The data supports the broader principle. Sanjay Gupta, a principal analyst at Forrester covering customer service operations, has noted that organizations still optimizing for handle time are “managing the symptom, not the system” — a framing that aligns with what Fapohunda observes in practice. 2026 industry benchmarks show operations that integrate journey design, AI, and analytics achieve around 23% higher CSAT and up to 30% lower costs. Operations focused on AHT in isolation achieve neither. The scorecard, Fapohunda notes, signals what leadership actually values. If handle time sits at the top, the organization will optimize for handle time — regardless of what the strategy documents say.
The Next 18 Months Will Separate the Operations That Got the Sequence Right
Gartner projects agentic AI will autonomously resolve roughly 80% of common service issues by 2029. McKinsey’s 2026 State of AI report finds that only around 30% of organizations have reached governance maturity level three or higher on agentic controls. Fapohunda sees that gap as precisely where the next wave of incidents and regulatory action will originate — and says he is building governance frameworks now, before the pressure arrives.
His approach relies on tiered autonomy calibrated to demonstrated reliability: fully autonomous for low-risk and reversible actions; human-approved for medium-risk or customer-impacting decisions; human-led with AI assistance for high-risk, regulated, or emotionally complex situations. The non-negotiable principle underlying all three tiers is reconstructability. “If my team cannot explain to a regulator, an auditor, or a customer why an agent did what it did,” Fapohunda says, “that agent should not be making that decision autonomously yet.”
Gartner projects 40% of agentic AI projects will be cancelled by the end of 2027. Fapohunda’s view is that the differentiator will not be ambition or budget. It will be sequence: governance first, journey redesign second, role redesign third, technology fourth. Inverting that order produces the incidents, the regulatory scrutiny, and the board conversations nobody wants.
The operations leaders who build the governance foundation now will deploy faster eighteen months from now. Those who invert the sequence will spend that same period explaining to regulators and boards why the technology moved faster than the controls around it. The window to get the sequence right is narrowing.
Follow David Fapohunda on LinkedIn for more insights on contact center operations, AI governance, and building operating models that deliver journey-level outcomes at scale.
