Skip to main content
DATUM
Enter
AI at Work Foundations · 0 of 8 modules complete
Back to LearnOpen Toolkit

Eight practical modules for work

One lesson, one lab, one call, one recipe. About two hours.

Module 1 · 15 min

What AI is and is not

Prediction, not a database: where language models help at work and where fluent guessing fails.

1. LessonSure but wrong
2. Sandbox labCompare two briefs side by side
3. Daily five callSpot an ungrounded claim
4. Recipe keptSummarise only from my pasted notes
Module 2 · 15 min

What can and cannot leave the building

Spot names, NHS numbers, payroll rows, and client secrets before anything hits a browser tab.

1. LessonWhat it cannot see
2. Sandbox labBefore you paste scrubber
3. Daily five callSafe to paste check
4. Recipe keptClean handover note with placeholders
Module 3 · 15 min

Writing a brief that holds up

Give the source, the reader, the format check, and what to say when a fact is missing.

1. LessonThe context window
2. Sandbox labBuild a six-field brief and test in Compare
3. Daily five callWrite the brief four-check rubric
4. Recipe keptSix-field work brief template
Module 4 · 15 min

Working from your own documents

Ask questions over a real policy or contract with verbatim quotes, and catch one planted mistake.

1. LessonRetrieval
2. Sandbox labDocs lab: quote check and planted mistake
3. Daily five callVerify every quote against the passage
4. Recipe keptAnswer from this document with exact quotes
Module 5 · 15 min

Asking a spreadsheet without trusting the sum

Run real table code over your rows and verify one cell by hand before sharing the figure.

1. LessonEvals
2. Sandbox labData lab: ask a spreadsheet and check one number
3. Daily five callFormula versus chat arithmetic
4. Recipe keptTable breakdown with one manual cell check
Module 6 · 15 min

Spotting a smooth wrong answer

Break a fluent AI reply into checkable claims and test whether it is simply agreeing with you.

1. LessonIt agrees with you
2. Sandbox labCheck an AI answer
3. Daily five callWhich one could you send?
4. Recipe keptAsk for the three weakest assumptions in my draft
Module 7 · 15 min

Scams, voice clones, and synthetic media

Apply the call-back rule on urgent voice notes, video requests, and synthetic charts or photos.

1. LessonPrecision and recall
2. Sandbox labCompare a real request with a urgency trap
3. Daily five callThe directory call-back rule
4. Recipe keptOut-of-band call-back check for urgent requests
Module 8 · 15 min

Rules at work: policy, copyright, bias, and when a person has to decide

Keep human approval gates on external actions, test summaries for name bias, and follow workplace policy.

1. LessonPrompt injection
2. Sandbox labAgent lab: hand off a task with approval gates
3. Daily five callHold external actions for a person
4. Recipe keptOne-page team AI policy and approval gate
Shareable Foundations record

Plain record of the eight modules and real labs completed

Lists the modules completed, the date, and the labs and checks behind them. No score, no percentile, and no rank. You can revoke the link at any time.