Course
All chapters

gooder. course · Chapter 3

Output: human review that means something

Everyone12 minutesUK edition
Video · about a minute. Vaughn walks through a real AI-drafted client email, line by line, and shows the four things a proper check catches.

Check AI output properly before anyone relies on it, and match the depth of review to the risk.

Why it matters

Your customers

Clients pay for your judgement. If they find an error AI made and you didn't catch, they don't blame the tool. They wonder what they're paying you for.

57%of UK consumers say they can usually tell when content is AI-generated. 66% think brands now put volume ahead of quality.Source: Optimizely, September 2026

Your bottom line

When AI-written work goes out unchecked, the cost lands on you: refunds, regulators and lost clients.

$290,000about how much of its fee Deloitte refunded after an unchecked AI-assisted report for a government client cited sources that didn't exist.Source: Fortune, October 2025

56% of people who use AI at work say it has caused mistakes in their work.

Source: KPMG and University of Melbourne, May 2025

Locked

The rest of this chapter is for Plus and Pro teams.

Plus and Pro include the full course for your whole team, with progress, checks and a training record for every person.

In this chapter

  • Review before reliance
  • What to check
  • Match the review to the risk
  • What makes review meaningful
  • Does it sound like you?
  • A step-by-step walkthrough to do today
  • 1 real case, with sources
  • What we'd do, in real situations
  • 3 questions people ask
  • A 4-question check, with a record when you pass

UK edition · Switch to the US edition