Five technology waves, one repeating pattern: what 26 years of transitions predict about your AI program

The technology changes every five years. The failure never does.

I’ve spent 26 years helping organizations cross technology transitions — from the early web to connected experiences, to omnichannel, to digital transformation, and now AI. European Parliament, Belgian public administrations, energy companies, telecoms, retailers. Different decades, different buzzwords, different vendors.

And every single time, the same pattern.

The pattern

It starts with excitement at the top. Leadership sees the wave coming, buys the vision, announces the program. Budgets appear. Steering committees form. A roadmap gets printed.

Then the gap opens: the gap between “we’ve decided to do this” and “people are actually doing it.” Teams have talent but no direction. Stakeholders each hold a different picture of the destination. Middle managers translate fog into more fog. And the traditional response — bigger decks, tighter governance, more framework — makes the complexity worse, not better.

The program doesn’t die loudly. It stalls quietly.

Five waves, five versions of the same lesson

The early web taught me that access is not adoption. Everyone rushed to have a website; almost nobody asked what users actually came to do. Years later at the European Parliament, we tested the site’s navigation and measured a success rate below 20%. Tens of thousands of pages, 27 languages, peak traffic of over a million visitors a day before elections — and four out of five visitors couldn’t complete their task. The technology was there. The clarity wasn’t. It took more than 50 stakeholder workshops and over 6,000 user tests, applying Top Tasks Analysis with its creator Gerry McGovern, to raise that success rate above 82%. What fixed it wasn’t more technology. It was ruthless clarity about what mattered.

The social and connected wave taught me that channels multiply faster than understanding. Organizations added platforms the way they once added pages — presence everywhere, purpose nowhere.

The omnichannel wave taught me that structure eats strategy. At Carrefour, the question wasn’t “should we do digital?” — everyone agreed. The question was which format serves which customer journey, from Express to hypermarket, from pickup to home delivery. The hard work was simplification: turning a wall of possibilities into a picture people could act on. The best compliment I received there still guides me: “In business we tend to create complexity — his greatest strength is that he simplifies things.”

The digital transformation wave taught me that products beat programs. At IRISnet, we replaced a maze of back-ends, interfaces and semi-manual processes — including an Excel file on a shared drive that qualified as a “system” — with one coherent digital product. Three years later it was the flagship: more than 50% of customers onboarded in the first year, NPS up 40 points. Not because we had a bigger transformation program. Because we gave people one clear thing that worked, and let momentum do the rest.

And the AI wave? It’s the same movie, at higher speed. Leadership buys the vision. Pilots multiply. And a 2026 survey by Writer found that 75% of executives admit their AI strategy is “more for show than actual guidance.” The tooling is unprecedented. The paralysis is not.

Why this keeps happening

Because organizations treat each wave as a technology problem, when it has always been a clarity problem.

Every wave produces the same three fears — becoming obsolete, losing control, being exposed as incompetent. Every wave produces the same misalignment: strategy people, technology people and operational people each holding a different map. And every wave punishes the same reflex: responding to uncertainty with governance instead of direction.

Transformation programs fail at rates estimated around 70% — and the recurring cause isn’t the technology. It’s human disengagement in the fog.

What this means for your AI program

If you’re leading an AI transition right now, the pattern predicts your next 18 months better than any vendor roadmap:

• The pilot phase will feel like progress. It isn’t, yet.
• The gap between decision and behavior will widen quietly.
• The people problem will surface later than the technology problem — and cost more.
• More governance will feel safer than more clarity. It won’t be.

The organizations that cross — I’ve watched them do it five times now — do something different: they invest first in a shared, concrete picture of where they’re going, make the path small and visible, and treat their people’s energy as the critical resource it is.

The wave is new. The bridge is not.

Sources

• Writer, Enterprise AI adoption 2026