Two exercises, 60 minutes total. The first is a quick recap. The second is the capstone: a cold customer conversation with no FitPath scaffolding to lean on.

Overview

• Learning objective: Consolidate how a single AI capability's architecture and governance requirements matured over the course — proof of retained judgment, not just retained facts.
• Time estimate: 10 minutes
• Difficulty: Low
• Setup / materials: Your own memory and notes from the course. No AI tool needed for this one.

Participant instructions

Pick one FitPath capability that showed up in more than one day of the course. Some candidates:

• Weekly workout plan generation (Days 1, 2, 4, 8)
• The Nutrition Agent / Workout Agent pair (Days 5, 7)
• The Trainer Assistant (Day 10)
• Anything else you tracked across multiple sessions

Write a short trace — five bullets, no more than a sentence or two each:

1. What it was. One line on what the capability actually did.

2. What changed technically between its first and most recent appearance.

3. What changed about the governance answer between its first and most recent appearance — be specific (a control that got added, a permission that got scoped, an approval gate that appeared).

4. What didn't change — the part of your answer that would have been correct even on Day 1.

5. One closing line: if Priya asked you about this capability today, what's different about your answer versus how you'd have answered on Day 1?

• You can name a specific architectural or governance shift — not "it got more secure," but what changed and why.
• Bullet 4 shows you can tell the difference between what you learned and what you already knew.
• The closing line demonstrates you're thinking about your own growth as an advisor, not just the system's evolution.

Extension

If you finish early, do a second capability and compare: did governance requirements mature the same way for both, or differently? Why?

Your comparison

Overview

• Learning objective: Run the SA Advisory Playbook end-to-end on a scenario you've never seen — no persona history, no prior architecture to reference, no narrative shortcuts. This is the transfer test: does the judgment generalize past FitPath?
• Time estimate: 50 minutes (5 min read-in · 30 min live conversation · 15 min written roadmap)
• Difficulty: High — this is the capstone exercise for the course.
• Setup / materials: Your AI tool of choice (Anthropic Console, ChatGPT, Google AI Studio). You'll paste in a starter prompt to run the stakeholder as a live conversation partner.

Participant instructions

You've just been looped into this inbound request. This is everything you know:

From: Dana Reyes, VP of Claims Operations, Meridian Mutual
Subject: AI for claims — need a plan

Meridian Mutual is a regional P&C insurer, roughly 2 million policyholders across the Midwest. Our COO wants claims processing time cut in half by year-end and has been told AI can get us there. Setting up 30 minutes to talk through what's possible.

That's it. No prior relationship, no course history, no persona notes to fall back on.

Your task, in order:

1. (5 min, silent) Re-read the brief. Jot down — on paper, not to the AI — what you don't yet know and what your first question will be. Do not open with a technology question.
2. (30 min, live) Paste the starter prompt below into your AI tool and run the conversation as Dana. Use the Strategic Customer Conversation structure from today's session: Business Ambition → Use Case Selection → AI Fit → Architecture Pattern → Governance, Value & Roadmap. Avoid the anti-pattern — don't lead with model, framework, or "agent vs. RAG" questions.
3. (15 min, written) Close your AI tool and write a one-page adoption roadmap memo addressed to Dana and her leadership. Structure it around the 7-phase roadmap (Discover → Optimize) and be explicit about governance requirements — this is an insurance company; the governance conversation will not look like FitPath's.

If Dana doesn't know something (data readiness, for instance), that's realistic — don't invent an answer for her. Capture what you don't know as an open risk in your roadmap.

Your one-page adoption roadmap memo

Paste this into your AI tool exactly as written to start the roleplay:

(Facilitator will provide the full behavioral brief separately — don't worry about it, just start talking to Dana once the prompt is loaded.)

• You opened with the business process/ambition question, not a technology question.
• You proactively probed Meridian's regulatory and compliance context rather than taking "cut time in half" at face value.
• If Dana proposed something with real regulatory weight, you caught it and raised the governance implications before she had to spell them out for you.
• If Dana asked for a fast, broad rollout, you proposed something more structured — and explained why.
• Your memo names at least one open risk or unknown rather than inventing certainty Dana didn't give you.
• Your memo maps cleanly onto the 7-phase roadmap and names concrete governance controls specific to an insurance claims context — not FitPath's governance controls with the names swapped.

Extension

Fast finishers: rerun the conversation, but this time Dana insists on full claims autonomy by next quarter and won't accept a phased answer. Defend a governance-first position without losing the business relationship. Save both memos and compare which one you're more confident presenting to a real customer.

Your comparison