FAST-09 Projects: AI for Managers and Leaders
Every project is built on your own team, sanitized: roles instead of names, no employee or customer personal information, no salaries, bids, or confidential figures. Your manager or the business owner should be comfortable reading anything you submit. Scoring: each criterion is Not yet / Meets / Strong; completion needs Meets on every capstone criterion (one resubmission allowed within the course window).
Capstone: 90-Day AI Adoption Plan for My Team
What you produce: a 2-3 page plan, plus three attachments from the labs: W2 use-case matrix, W3 ROI worksheet, W4 team AI use guideline. Peer-reviewed in session 3 and scored by the instructor.
Required sections
| Section | Content |
|---|---|
| Team and starting point | Team size and roles (no names), current AI use (including unofficial use), employer policy status |
| Chosen pilots | Two use cases from W2, with value and risk reasons; what you parked and why |
| Business case | W3 math for pilot 1: hours saved incl. checking × loaded cost × weeks, minus tool, training, maintenance; assumptions with sources; weakest assumption |
| Guideline | W4 attached; how it fits the employer policy and Massachusetts data rules |
| Phases | Days 1-30, 31-60, 61-90: milestones, owner, date |
| Training | Who, what, when; before anyone is expected to use AI |
| Measures | Baseline and target per pilot; team-level only; how results are shared |
| Communication | Kickoff message, cadence, how questions and concerns are handled |
| Risks and stop conditions | Top three risks, the control for each, and what would make you stop a pilot |
Steps
| Step | When | What you do |
|---|---|---|
| 1 | After session 1 | List 5 places AI might help your team (activity F09-S1-A5) |
| 2 | Session 2 | Score 8-10 candidates (W2); ROI for your top pick (W3) |
| 3 | Session 3 | Write W4; draft the plan; peer review in pairs |
| 4 | Within 7 days | Revise using peer and instructor notes; submit |
Rubric
| Criterion | Not yet | Meets | Strong |
|---|---|---|---|
| Judgment | Pilots chosen without reasons, or a high-risk use case first | Two pilots with sound value and risk reasons; high-risk ideas parked | Data and policy questions identified and assigned before the pilot starts |
| Business case | Math wrong or checking time left out | Correct math; checking, training, and tool costs included; assumptions stated | Sensitivity shown (what if checking takes twice as long?) |
| Policy fit | Guideline conflicts with employer policy or allows restricted data | Guideline consistent with employer policy and G.L. c. 93H / 201 CMR 17.00 | Reporting path for data slips tied to the employer's security program |
| Plan quality | Phases, training, or measures missing | Milestones, owners, training, baseline and target, communication present | Stop conditions and a keep/change/stop decision meeting scheduled |
| People | Monitoring of individuals, or no plan for concerns | Team-level measures; honest message; time for questions | Staff invited to propose and test use cases |
Weak vs. strong (one line each)
| Weak | Strong |
|---|---|
| "We'll roll out AI to everyone in January." | "Days 1-30: guideline Nov 2; 45-minute training Nov 5; pilot 1 (guest email replies from approved FAQ) with 4 front-desk staff. Baseline 6.5 h/week (2 weeks of logs); target 3.5 h/week with zero wrong policy statements sent." |
| "AI will save us a lot of money." | "Net first-year estimate $9,659: 4.67 h/week × $46.89 × 48 weeks, minus $281 training and $563 template maintenance. Weakest assumption: 5 min per letter including checking, re-measured in days 31-60." |
| "Managers will keep an eye on usage." | "We measure team minutes per task and errors caught in review; staff self-report weekly in two lines; results shared at the week-6 huddle." |
Mini project 1: Vendor AI Feature Decision Memo
Scenario: A vendor (fictional) offers to add an AI feature to software your team already uses. You write a one-page buy / build / wait memo for the owner.
Provided (fictional): ClaimDraft Pro proposal excerpt from Lab 1, or any public product page for a tool your employer already licenses (no confidential pricing).
Memo contents: the task it would help with; expected benefit (from W3-style math); what the terms say about price, term, auto-renewal, price increases, data use for vendor improvement, storage location, and support; questions to send to the vendor in writing; your recommendation (buy, build, or wait) and the condition that would change it.
| Criterion | Not yet | Meets | Strong |
|---|---|---|---|
| Terms read | Price only | Price, term, renewal, data use, and support all captured | Cancellation deadline and data opt-out turned into dated action items |
| Benefit | Vendor's claim repeated | Own estimate with checking time | Range (low/likely/high) with assumptions |
| Recommendation | No clear decision | Clear buy/build/wait with reasons | Names the evidence that would change the decision |
Example recommendation line: "Wait: the features we need already exist in our current suite; the proposal auto-renews with up to 9% increases and uses our content for vendor improvement unless we opt out. Revisit if the 30-day suite test shows less than 2 hours a week saved."
Mini project 2: The Conversation Kit
Scenario: You will introduce AI to a team where some people are worried. Build a kit you can actually use.
Contents: (1) a 150-word kickoff message; (2) answers to five likely questions (for example, "Will this replace me?", "Will you read my prompts?", "What if I make a mistake?", "Do I have to use it?", "Who decides what's allowed?"); (3) a 10-minute huddle agenda; (4) a two-question weekly self-report form; (5) one paragraph on how you will collect and act on staff ideas.
| Criterion | Not yet | Meets | Strong |
|---|---|---|---|
| Honesty | Promises nothing will change, or threatens | States what changes and what doesn't; no promises you can't keep | Explains how AI-assisted work will be judged |
| Respect for privacy | Implies reading prompts or individual tracking | Team-level measures; self-report only | Explains what the employer's tools can and cannot see |
| Usability | Generic text | Ready to use next week, fits your team | Tested with one trusted team member and revised |
Example answer to "Will you read my prompts?": "No. We'll measure how long the task takes the team and how many errors we catch, and you'll tell me in two lines each week what helped and what didn't. Our IT team manages the tool under the company's policy; if you have a question about what the system records, ask me and I'll get you a straight answer."