How to Run an AI Pilot That Doesn’t Die

By Johannes Sundlo — AI & Future of Work Advisor. I help leaders and HR teams turn AI into adoption that sticks, through keynotes, workshops and change programs.

Most AI pilots don’t fail because the technology was wrong. They quietly die because they were set up to prove a tool instead of build a habit. Here’s how to run an AI pilot that actually leads somewhere.

Why pilots quietly die

• They’re disconnected from real, daily work — a demo, not a workflow.
• There’s no owner with the time and mandate to carry it.
• Success was never defined, so nobody can say if it worked.
• The first rough results get read as “AI doesn’t work here.”

A pilot that leads to adoption — step by step

1. Pick a real, repeated task. High-frequency, low-risk, and genuinely annoying to a specific team. Frequency compounds the payoff; low risk lets people learn without fear.
2. Name an owner and give them air cover. One person accountable, with protected time and a leader shielding them from ‘prove the ROI by Friday’.
3. Define success up front. Pick 2–3 measures: time saved, usage without being told, and whether quality holds. Write them down before you start.
4. Build it into the workflow. Put the AI step inside the existing template or checklist — not a separate tool.
5. Run it for a few weeks, then decide. Keep, kill, or expand based on the measures — not on vibes or politics.
6. Tell the story. Share the concrete result widely. That’s what turns one pilot into organisation-wide pull.

The four-week pilot plan

Four weeks is long enough to build a habit and short enough to keep energy. A structure that works:

• Week 1, set up. Owner writes the prompt or workflow with the team, defines the 2 to 3 success measures, and runs a 45-minute kickoff where everyone tries it on a real task. Nobody leaves without having used it once.
• Week 2 to 3, run. The AI step lives inside the normal template or checklist. Owner checks in twice a week, collects friction points, and fixes the prompt rather than the people.
• Week 4, decide. Compare against the measures you wrote down. Keep, kill or expand. Then, whatever the outcome, present it to the wider team: what was tried, what the numbers said, what happens next.

A mini-case: interview debriefs, with numbers

A composite of pilots I have run with talent and HR teams, with typical numbers. A team of six recruiters spends 20 to 30 minutes after every interview turning messy notes into a structured debrief. That is the pilot task: high-frequency (30+ interviews a week across the team), low-risk (a human reviews every summary), and genuinely disliked.

Setup: one prompt that takes raw notes and returns the team’s scorecard format. Success measures written down up front: time per debrief, voluntary usage in week three, and whether hiring managers notice a quality drop. Typical outcome after four weeks: time per debrief falls from about 25 minutes to under 10, five of six recruiters use it without being reminded, and quality holds because the recruiter edits instead of writes. That is roughly 10 hours a week back for one team, from one prompt, and, more importantly, six people who now trust AI on a real task and start asking “what else can we point this at?” That question is the actual return on a pilot.

How to know your pilot is working

The clearest signal isn’t a glowing survey — it’s people using the AI step without being reminded, and the human editing getting lighter over time. When usage spreads on its own, you’ve got real adoption, not a demo.

A pilot is step one. For the full system around it, see the hub guide on AI adoption that sticks, plus AI for Leaders on backing pilots from the top.

Common questions about AI pilots

How long should an AI pilot run?

Four weeks is the sweet spot: long enough for a habit to form and for early roughness to wear off, short enough that energy and sponsorship survive. Pilots that run “until we know” run forever.

How many pilots should we run at once?

One or two, not ten. Every pilot needs an owner with real time, and attention is the scarce resource. One pilot that lands and gets talked about creates more organisational pull than ten that limp.

What is a good first pilot task?

Something high-frequency, low-risk and disliked: meeting or interview summaries, first drafts of routine documents, structuring free-text feedback. Avoid anything where an error is expensive or a decision about a person is involved.