The tool is not the transformation: what finally made AI stick on my team

As a tech leader, I’ve seen countless tool rollouts; some stick, most fade away. When we introduced AI into our development workflow (specifically, Claude, Cursor and a few other tools) I anticipated some friction, but I assumed the sheer potential would win everyone over.

It didn’t.

Despite the hype, adoption was inconsistent. A few engineers adopted it, but many others refused to touch it. When leadership asked, “How’s the AI initiative going?” I lacked a straightforward narrative on its value.

That’s when I realised the fundamental truth: this was a change management challenge.

Where we went wrong: the myth of “build it and they will come”

In hindsight, our initial mistakes are common to many organisations:

• We provided access but failed to demonstrate a clear path to a “win”;
• We assumed our sharpest engineers would simply “figure it out” on their own time;
• We positioned it as a top-down mandate, not a team-led opportunity to solve shared problems.

The result was sporadic usage, quiet opt-outs, and no measurable business impact. We had an AI budget, but not an AI capability.

The breakthrough: credibility is built on solved problems

The turning point wasn’t a new platform or a company-wide mandate. It was solving one of our team’s most tedious tasks: writing QA documentation.

Using AI, I reduced a four-hour manual process to a 30-minute review. I shared the workflow in our next team meeting, not as a pitch, but as a practical solution to a shared headache. The reaction was one of curiosity rather than resistance.

From that single, tangible win, adoption grew organically. One engineer asked for the prompt. Another saw an opportunity to automate part of their code review prep. The growth was steady, authentic, and built on trust.

This idea of a small, practical win unlocking a bigger change isn’t unique to AI. I was reminded of the same pattern on a different project, where solving a tricky analytics module problem ended up supercharging our entire agile process. 

From compliance to ownership: the new normal

Today, our team’s relationship with AI is characterised by shared ownership, rather than compliance. Engineers proactively suggest improvements to AI workflows because they see the direct benefit.

When a new developer joins, they are onboarded with the same prompt templates, style conventions, and context-aware tools our senior team uses. Their output aligns with our quality standards from day one. AI enforces structure, enabling us to transition from chaotic sprints to the more deliberate, robust processes we described in The Agile Spike of the Future.

A playbook for tech leaders

If you’re starting this journey, my advice is to trade grand gestures for small, credible wins.

You have to be the first user, identify a painful, universally recognised problem within your team and solve it yourself using the new tools. Let that success serve as a beacon for sceptics, demonstrating the undeniable value of your team’s daily work.

The ultimate goal is to answer the critical question: How do we integrate AI into our culture, not just our technology stack?

My most important lesson is this: you cannot delegate AI transformation. It’s not a procurement task you can hand off, it requires leadership.

That means showing up, experimenting personally, and converting small wins into unstoppable momentum. When you lead from the front, AI stops being a buzzword on a slide deck and starts becoming a real, compounding advantage for your team and your business.