AI doesn't fail because it's too ambitious. It fails because it's unfocused. In my last blog, I unpacked the risk of racing into AI initiatives that look impressive in slide decks, but deliver little real value for customers or the business. The truth is simple: solving two critical pain points will always outperform delivering ten quick wins. Let me walk you through how to find yours.
Find Points of Customer Friction
Think about the interactions that create the greatest emotional burden for customers and the heaviest operational load for your teams. These are high-cost, multi-touch interactions that generate friction, frustration and become the stories customers repeat to family and friends. Left unaddressed, they quietly erode your brand.
A good place to start your investigation is with your top complaint drivers. Pull the verbatim comments because there is real power in the words customers use. Then, layer in the supporting data: repeat call rates, handle times, escalation patterns. Next, add a source that's closer than you might think: your customer-facing teams. Ask them what gets in the way of delivering exceptional service. You'll get an unfiltered view of where policy, process or systems break down in real time. No summary report or dashboard captures that. Take all of it together, map the related processes to identify root causes, and trace it through your end-to-end complaint handling process. What you'll uncover is revealing: where the biggest pain points truly live, how quickly complaints are resolved and the real cost of service recovery.
Something that has always stuck with me is the research by TARP (Technical Assistance Research Programs at Harvard) who found that customers whose complaints were resolved quickly showed loyalty rates above 95%. That’s higher than customers who never experienced a problem at all. A well-handled complaint doesn’t just repair the relationship; it can actually strengthen it.
Speed, however, is only part of the story. Slow resolution multiplies cost through repeated contacts, compensation and reputational damage. According to a Nielsen-McKinsey study, only 17% of customers would recommend a brand that resolves issues slowly, even when the outcome is ultimately successful. Poor complaint handling doesn’t just create a momentary failure; it leaves a lasting mark by damaging brand reputation, eroding customer trust and causing customers to walk out the door.
Shift to Shared Accountability
Identifying the problem gets you halfway there. I was in a room with a leadership team last year that had just launched their third AI initiative in 18 months and still couldn't articulate the ROI of the first initiative. That's not an AI failure. That's a diagnostic and accountability failure.
The answer is shared ownership. Bring together a cross-functional SWAT team: Marketing, Sales, Operations, Technology, HR. Seat them at the same table with shared KPIs that make collective accountability the expectation, not the exception.
Let me give you a real example of what this looks like in practice. We assembled a cross-functional team to implement an AI chatbot with two objectives: improve customer experience and reduce operational cost. When results fell short of the business case, the team didn’t point fingers. They identified root causes, designed corrective actions and delegated execution across responsible departments. At the end of the proof of concept, they conducted a formal postmortem. The mindset was simple: when one team fails, everyone fails.
What makes this different from traditional continuous improvement efforts is integration. Continuous improvement initiatives often stall because they’re owned by Operations alone and struggle to compete with other departmental priorities. The cross functional team model works because accountability and KPIs are agreed upfront. Findings don’t become a laundry list of issues at an executive meeting. The top two become the focus. And the work doesn’t stop after one fix. The team cycles back and plans the next solution. That’s how strategic planning becomes a continuous improvement engine, not a one-time exercise.
The Compounding Effect of Working Together
Fix the right two things and the impact compounds in ways that matter. You reduce customer frustration and build trust, remove the interactions that emotionally drain your teams and shift the cost of the entire customer journey, not just shave seconds from a call.
The most powerful strategic plan isn't the most ambitious. It's the one that starts with the right diagnosis. Most organizations already have the data. The discipline is deciding what to do with it and holding everyone accountable for the outcome.
In my next blog, I'll explore how agentic AI fits into this, and how to decide whether it's the right prescription for your top two pain points.