Book a chat
Work Feed Guides Services Approach About Contact Book a chat
Edition 02 · Free PDF · Nurture

The 30-Day AI Pilot

A day-by-day guide to running your first AI pilot — what to measure, when to stop, and how to know if it worked.

Get the free guide See what's inside
12 pages 12 min read A4 print-ready Free PDF

Get the free guide.

You've decided AI might help. Now what? This 12-page guide gives you a day-by-day plan for a 30-day pilot: what to set up in week one, what to measure in week two, how to read the results in week three, and how to make the go/no-go decision in week four. Includes a stop criterion — because a pilot that ends in 'not yet' is a successful pilot.

We'll email you the download link. No spam, no newsletter — just the guide.

What's inside

6 sections, 12 pages.

  1. Why 30 days, not 30 minutesA pilot that lasts 30 minutes tells you nothing. A pilot that lasts 30 days tells you everything. The guide explains why the middle ground is a trap.
  2. Week 1: set up the baselineBefore you touch AI, measure how long the task takes today. Without a baseline, you can't prove the pilot worked — and you can't prove it failed.
  3. Week 2: run the pilot alongside the old wayThe old way and the new way run side by side. Same task, same person, two methods. The comparison is the whole point — and it's easier than it sounds.
  4. Week 3: read the numbersThree numbers to track: time saved, error rate, and whether the person doing the task actually trusts the output. The guide gives you a one-page tracking sheet.
  5. Week 4: the go/no-go decisionThe decision isn't 'did it work?' It's 'did it work enough, reliably enough, for the cost?' The go/no-go matrix turns gut feel into a defensible answer.
  6. The stop criterionA pilot that ends in 'not yet' is a successful pilot. The stop criterion saves you from six months of limping along with a tool that isn't ready.
Common questions

Questions this guide answers.

How do I run an AI pilot in my business?
Pick one task. Run the AI system alongside the existing process for 30 days. Measure time saved, error rate, and team acceptance. The guide gives you a day-by-day plan: week one is setup and baseline measurement, week two is parallel running, week three is data review, week four is the decision. The key rule: never replace the old process until the pilot proves the new one is at least as good.
How long should an AI pilot last?
Thirty days is the sweet spot for most UK SME tasks. Shorter than two weeks and you don't have enough data to trust the results. Longer than six weeks and you lose momentum. The guide explains why 30 days works and what to measure each week.
What should I measure in an AI pilot?
Three things: time saved per task (compared to the baseline), error rate (how often the AI output needs human correction), and team acceptance (does the person who used to do the task trust the AI's output?). The guide provides a simple tracking sheet and explains how to read the numbers without over-interpreting early results.
When should I stop an AI pilot?
Stop if the error rate is above 20% after two weeks of tuning, if the team actively resists using it, or if the time saved is less than the time spent reviewing the AI's output. The guide includes a stop criterion worksheet — because a pilot that ends in 'not yet' teaches you more than one that limps along for six months.

Other guides in the series

01Where AI Actually Helps 02The 30-Day AI Pilot (this guide) 03Your Data, Your Rules 04Writing AI Instructions That Actually Work 05The AI Vendor Interrogation Kit 06Your First AI Employee 07Set Up Your Own AI Agent 08Is This Worth Doing? 09Build the Thing 10Get Found 11Get the Conversation 12Close the Deal

Read the guide. Then talk to us if you want to.

We'll tell you on the first call if the guide is all you need.

Book a chat