How to Become Your Company's AI Champion in 90 Days

Every organisation is developing an unofficial AI hierarchy right now. There's usually one person everyone quietly asks — not the most senior, not necessarily technical. Just the person who figured it out first and was generous about sharing.

That role is worth having. It changes how you're perceived, what you're included in, and what happens to you in a restructure. Here's how to occupy it deliberately in about three months.

Days 1–30: Build Real Competence Quietly

The failure mode here is announcing yourself before you have anything to offer. Spend the first month building something to be credible about.

Week 1–2: Get properly fluent at briefing

Take three tasks you do weekly and rebuild them with AI. Not new tasks — familiar ones, so you can judge output quality. Save the prompts that work.

Week 3: Map where AI fits in your own function

List every recurring task in your role involving summarising, drafting, comparing, formatting or organising. Mark each one: AI does this well / AI helps partially / AI shouldn't touch this. That third category matters — being the person who knows the limits builds more trust than evangelism.

Week 4: Build one thing worth showing

One workflow that saves real, measurable time. A reporting process that took four hours and now takes forty minutes. Something concrete you can demonstrate in three minutes.

At the end of month one you should be able to finish this sentence: "I cut [specific task] from [X] to [Y] using [approach]." That sentence is your entire credibility. Everything else builds on it.

Days 31–60: Make Yourself Useful to Others

Now you have something, start giving it away. Generosity is the whole strategy — internal influence follows usefulness, not expertise.

Week 5–6: Help three individuals

Pick three colleagues in different functions. Ask what part of their week they find most tedious. Then actually solve it with them — sit down, build it together, leave them with something working.

Three people who tell others you helped them is worth more than any presentation.

Week 7: Write the internal guide nobody has written

A short document: which tools your organisation permits, what should never be pasted into a public AI tool, five prompts that work for common internal tasks, and who to ask.

Every organisation needs this and almost none have it. Writing it makes you the de facto owner of the topic, without ever claiming the title.

Week 8: Run one low-stakes session

Thirty minutes, one team, one specific use case relevant to them. Not "an introduction to AI." Something like "how to cut your monthly report prep in half." Demonstrate, let them try it live, leave.

Days 61–90: Convert Goodwill Into Position

Week 9–10: Gather evidence

Go back to everyone you helped and ask what changed. Collect specifics: hours saved, turnaround improved, error rates down. You need numbers, because the next step requires them.

Week 11: Take a proposal upward

One page to a decision-maker. Structure it as: what we're currently losing, what three people already fixed and what it saved, what a wider rollout would need, and what you're asking for.

Ask for something modest and specific — budget for a team programme, permission to run monthly sessions, an official remit for a few hours a month. Vague requests get vague answers.

Week 12: Make it structural

Whatever you got approved, give it a shape that survives you being busy: a recurring session, a maintained internal doc, a named group. Informal influence evaporates when you have a hard quarter. Structure doesn't.

The Four Ways Internal Champions Fail

Worth knowing in advance, because these are predictable.

1. Evangelising before delivering

The person who talks about AI constantly but has never shown a working result becomes background noise within a month. Build first, talk second.

2. Making colleagues feel stupid

The fastest way to kill adoption is to make people feel behind. Frame everything as "here's a shortcut I found" rather than "here's what you should be doing." Nobody adopts advice from someone who makes them feel slow.

3. Ignoring the genuine risks

Dismissing data security or accuracy concerns marks you as naive and gives opponents an easy target. Being the person who says "don't put client data in that tool, here's what to use instead" makes you more credible, not less.

4. Threatening someone's territory

If IT owns tooling decisions or L&D owns training, going around them creates an enemy with a legitimate grievance. Bring them in early. Shared credit costs you nothing and buys you the ability to act.

What This Is Actually Worth

Being your organisation's AI champion means you're in conversations about how work gets redesigned rather than being told the outcome. In a period where a lot of roles are being reshaped, that's the difference between shaping the change and receiving it.

It also compounds externally. "Led AI adoption at [company]" is currently one of the more valuable lines on a CV in this market.

Where to Start This Week

Establish your baseline honestly with our free AI Readiness Score — ten minutes, behavioural questions, role-specific output.

If you want the 90-day path compressed into four structured weeks with a cohort and feedback, that's precisely what our AI Champion programme is built to do.