Twenty years of rolling out CX improvement programs has taught me one unglamorous truth: the strategy is rarely the problem. I’ve seen beautiful journey maps, watertight business cases and dashboards that would make a data scientist weep, then watched them quietly shelved a few months after launch. Not because the ideas were wrong, but because becoming customer-centric asks people to work differently, and nobody managed that part.
If you lead CX, you’ll recognise this. And if your organisation is now rushing to “add AI”, you’re about to watch the same pattern play out again, only faster and at greater cost.
Peter Drucker put the why simply in 1954: the purpose of a business is to create a customer. Theodore Levitt later added the part CX people care about most: and keep them. Simple to say. But living it means asking your Finance Manager to weigh customer lifetime value against this quarter’s numbers, or telling your Product Team to stop building a feature customers never wanted. Improving CX isn’t a project you bolt onto the side of the business. It’s a change in how people think, decide and behave, from the contact centre to the boardroom.
Over 20 years I’ve come to rely on two things to make that change stick: urgency and measurement. Get those right and the rest follows. Get them wrong and even the best strategy dies in a drawer.
1. Build the urgency
John Kotter, who has studied organisational change for decades, found that more than half of the transformation efforts he observed failed at the very first step: establishing a sense of urgency. Most leaders skip it. They assume the case is self-evident. It isn’t. Not to the contact centre agent whose bonus depends on average handle time, and not to the manager who’s watched three transformation programs come and go.
You have to make the change matter to people in their language, at their desk, before you ask them to work differently. Make your burning platform clear to everyone.
2. Measure the change, not just the outcome
Every CX leader has a dashboard. Fewer have the right one. An NPS score on the wall tells you the temperature; it doesn’t tell you whether behaviour is actually changing. So measure the change itself. Are staff using the new process? Are leaders making different calls? Are customers noticing?
Pair your lagging indicators (satisfaction, retention, revenue) with leading ones (adoption, effort, engagement). The leading indicators are your early warning system. And if a metric never changes a decision, take it out of your measurement program.
Why this matters more than ever: AI
Every organisation I talk to wants to “add AI”. Very few are treating it for what it is: the largest change management challenge they’ll ever face.
The early numbers make the point. MIT’s Project NANDA found that around 95% of organisations were seeing no measurable return from their generative AI pilots. Not because the technology was poor, but because AI was run as a technology rollout instead of a change program: no urgency built, and nothing measured beyond “we deployed it.” Meanwhile, the same research found that workers at more than 90% of companies were already using their own AI tools at work, sanctioned or not. The change was already happening. Leadership just wasn’t leading it or measuring it.
That’s not a coincidence, and it isn’t a new problem wearing a new label. It’s the same failure pattern I’ve watched derail CX programs for two decades, with the same two things missing, at greater scale and higher stakes.
Here’s the good news for CX leaders: if you’ve spent years driving adoption and behaviour change, you already have the playbook. You may be better placed to lead AI adoption than anyone in IT.
The CX playbook for AI
- Build the urgency. Why now, why us, and what happens if we don’t?
- Lead the culture; don’t just license the software. Adoption is a behaviour, not a login.
- Give people concrete, valuable use cases. Three to five that create value on day one.
- Measure the change. Adoption, hours saved, decisions improved and, critically, customer outcomes.
Because AI, like customer-centricity, isn’t something you install. It’s something people have to adopt, trust and weave into how they work. That’s change, and change is a discipline, not an accident.
Not sure where your organisation stands? Take the free 4-minute CX maturity check and see how ready you are to lead change.
A version of this article first appeared on Considered.