While most organisations are deploying predictive, generative, and agentic AI, few can directly connect those investments to revenue, customer outcomes, or profitability. As AI becomes cheaper and easier to build, the quality of the problem being solved becomes the primary determinant of value. Teams that measure success by releases, features, or utilisation risk falling into the “build trap”, where activity increases but business impact does not.

Define Value Before Building

Leading organisations are shifting their focus upstream, from building solutions to diagnosing problems. They start by identifying where value is leaking across customer journeys and operational processes, whether through unnecessary effort, delay, cost, or risk.

This includes steps that already destroy value today or appear effective today but will erode value over time. For customers, leakage often shows up as unnecessary effort, time, or cost. For firms, it appears in activities that add minimal or no customer value but still consume time, money, capacity, or quality.

This reflects a broader product leadership principle that sounds familiar: customers rarely care about features; they care about outcomes. Customers don’t wake up wanting an AI assistant, an agent, or a new feature. They want a faster resolution, a better decision, a simpler process, or a more predictable outcome. AI creates value only when it changes the economics of an experience, a workflow, or a decision.

Obsess About The Customer Need, Business Problem And Root Cause

As Cat Wu, Head of Product at Claude Code, argues, product teams must start by answering three questions: “Who are we building for, what problems are we trying to solve, and what are the top use cases?” The most effective AI products begin with a clearly defined problem, not a technology looking for a use case.

Getting to the root cause of the source of friction is key. Solving the wrong problem more efficiently still delivers the wrong outcome. Take Starbucks for instance, to address long wait times, the company initially deployed the Siren System to speed up drink preparation; however, the root cause wasn’t how fast baristas made drinks — it was how they sequenced orders.

What Digital And Product Leaders Should Do

The organisations that outperform in the next phase of AI adoption will not be those that build the most AI. They will be those that consistently identify high-value problems, understand their root causes, and apply AI where it can create measurable outcomes. Leaders should:

  • Define outcomes before funding AI. Every initiative should have a clear link to a business metric, customer outcome, or operational improvement.
  • Prioritise value leakage over technology opportunity. Focus on where customers, employees, and processes lose time, money, trust, or productivity.
  • Insist on root-cause analysis. Use techniques such as the Five Whys and fishbone analysis before selecting a solution.
  • Measure impact, not activity. Adoption and utilisation are useful signals, but revenue growth, cost reduction, retention, productivity, and customer experience should be the ultimate scorecard.

Access the problem definition template for building AI products that solve real problems in my latest report (client access only).

Share