# Quick-Reference Card: AI for Managers and Leaders

<p class="meta">FAST-09 AI for Managers and Leaders · 7formations · Instructor: Wadood Chaudhary · All examples fictional · Training, not legal advice</p>

## Your data line

Never in a prompt: employee records, reviews, discipline, pay, health notes, customer personal information (G.L. c. 93H, 201 CMR 17.00), PHI, bids, confidential drawings. Sanitize: roles, not names; no IDs.

## Judge every AI output

**Accuracy** (facts, numbers, dates, commitments) · **Fairness** (wording that disadvantages anyone) · **Tone** (right for the audience and news).

## Supervising AI-assisted work

"Which parts were AI-assisted, and how did you check them?" The sender is accountable. Coach a check; don't ban quietly.

## Use-case matrix

Score Value, Risk, Data sensitivity, Effort (1-5). Pilot high value / low risk. Park high-risk until controls exist. Decisions about people stay with people.

## ROI

(old h − new h − checking h) × loaded cost × weeks − tool − training − maintenance. State every assumption and source; name the weakest. Loaded cost fallback: BLS private industry $46.89/h (June 2026).

## Buy, build, or wait

Use what you already pay for first. Build only simple, low-risk helpers. Waiting is valid. Read the terms: price, term, auto-renewal, increases, data use and opt-out.

## One-page guideline

Approved tools · Never allowed · Encouraged / Allowed with a check / Off limits · Who checks · Reporting mistakes (blame-free) · Review date.

## People

Be honest, listen first, train before use, invite ideas. Measure outcomes, not people: team-level and self-reported.

## 90 days

1-30 guideline, training, pilot 1 · 31-60 measure, adjust, pilot 2 · 61-90 review, keep/change/stop.