Generative AI at Work / Unit 1 GAI 019 review, 2 to 3 minutes
One Task, Start to Finish
The idea
A worked example, constructed for this course, runs the whole pattern on an ordinary task: turning a new company policy into a one-page summary that staff will actually read.
In practice
The task falls to a team lead, who must summarize a twelve-page travel policy for her staff. First, the fit. A summary for a known audience is a strong task for generative AI, and the policy is internal but holds no personal data, so her company's approved tool is the right place for it. Second, the draft. She pastes the full policy text instead of naming it, says the readers are staff who travel twice a year, and asks for the five rules they most need, in plain language, under three hundred words. Third, the check. She lists the draft's claims and traces each one back to the policy. Four hold. One does not: the draft says receipts are required above fifty dollars, while the policy says twenty-five. She also notices that the draft left out the approval step for international travel, which the policy treats as mandatory. Fourth, the decision. She corrects the threshold, adds the missing approval step, adds a line saying the summary was drafted with AI and checked against the policy, and sends it under her own name with a link to the full text.
Where people slip
The error was small and fluent, the kind a reader skimming for tone would pass. The check caught it because she traced each claim to the source instead of judging the draft by how it read. The omission was caught only because she knew the policy well enough to notice what was missing. The tool saved her the writing. It did not save her the reading.
Your place on the trail
Every lesson rests on earlier ones and opens later ones. Tap any to jump.
Builds on
Unlocks
- Nothing yet depends on this.
Something wrong in this lesson? Flag it · Sources for Generative AI at Work