"It's not going to be by developing courses. It's going to be those human-centered skills." — Patti Ouzounian

The real work in the room wasn't abstract. L&D professionals and people leaders were trying to reframe a moment of genuine disruption — AI reshaping job markets, role definitions, and hiring expectations — into a concrete career opportunity they could act on before the window closed. The work was getting stuck in the same place for most people: they knew their value was shifting, but they didn't have a repeatable method for mapping what to protect, what to let go, and how to communicate what they bring now. That's what this live in-person working exchange was built to move forward.

Who was in the room:

Andrea Nunez (Discover/Capital One), serving as the Internal Client — anchoring the challenge in the real business pressure L&D professionals face when AI disrupts both their work and their market position. Mike Hoyt (ELE), as the AI Strategist — guiding the technical workflow live, step by step. Patti Ouzounian (ELE), as the Real-World Skeptic — testing what would actually work, not just what sounded promising. Alongside them: a room of practitioners from L&D, talent management, OD, and knowledge management — some in active transition, some navigating rapid internal scope expansion, some trying to build an internal business case for the first time.

Designing the CRIT Prompt Loop

Most people in the room had used AI for routine tasks. What the lab exposed was the gap between using AI and briefing it well.

The framework at the center of the working exchange was the CRIT Model — from Geoff Woods' The AI-Driven Leader — distributed as a physical card in the room. Four components: Context (give AI everything about your role, function, team, industry, company), Role (tell AI it's your career advisor and playbook builder), Interview (instruct AI to ask you exactly three clarifying questions before it builds anything), Task (specify the deliverables — a 30/60/90-day action plan, a stakeholder map, objection handling).

Mike Hoyt walked the room through each component live, then made the critical distinction: the only section most people needed to customize was Context. And context, it turned out, was where most people were underinvesting.

"The more specific you are about your role and function, the more effective this playbook will be." — Mike Hoyt

The proof came from the tables. One participant — Natalie — had her actual job description on hand and pasted it directly into the prompt. Her playbook came back precise and usable on the first pass. Others who described their role in general terms got general output and had to restart. The room named it cleanly during share-out:

"If you're not heavy on context, you have to go back." —

The Interview step — the "I" in CRIT, and the step Mike noted most people skip — turned out to be where the real diagnostic work happened. AI's three follow-up questions didn't just fill gaps. They surfaced things participants hadn't consciously mapped. One practitioner shared two of the questions her AI generated. The second one stopped the room:

"Where are you currently spending time that feels necessary but you suspect is not actually where you create your highest value?" —

That's not a course-building question. It's a career architecture question — and it came from the prompt, not the facilitators.

Safeguarding Highest-Value Energy

The 30/60/90-day playbooks that came out of the live exercise weren't all pointing in the same direction — and that was the signal.

An OD consultant in the room described what her playbook reframe actually looked like: "It's not how do I grow my consulting business — it's how do I design a business that protects and amplifies my highest value energy." The shift she named was from change management executor to organizational redesign architect in the era of AI. Bigger market position. Cleaner scope. Different conversation with clients.

Another participant described a role with significant legacy creep — doing operations work inside HR, manual processes that had accumulated over time. Her playbook gave her the 30/60/90 structure she needed to make the business case for what belongs where. Not a career development document. A stakeholder presentation with a path to executive sign-off.

A third came in planning to use the playbook for an external job search. She left thinking about how to use it to pitch her CHRO on an internal AI initiative — building the business case, anticipating objections, framing the ROI.

Same tool. Three completely different applications. All grounded in the same core move: stop describing what you do and start protecting what only you can do.

Beyond the Prompt: Building the Personal Agent

Patti Ouzounian closed the working exchange with a move most of the room hadn't considered yet.

After the previous Design Charrette in this series, she'd taken the playbook output and done something with it: she built a custom ChatGPT agent. Not a new chat. An agent — with persistent instructions, a defined role, and an automated task running in the background.

"I have it scanning 35 to 40 job posting sites — it ranks positions according to the descriptors I put in." — Mike Hoyt

The agent monitors the market continuously, ranks opportunities against explicit priority descriptors, and surfaces alignment without manual searching. The setup, it turned out, didn't require a tutorial. Mike Hoyt framed it: "I didn't need a video. I just had to follow the instructions. It's like a recipe."

The room immediately surfaced the next question: what makes some agents work well and others not? Same answer as the playbook — precision of the brief. Vague instructions produce vague agents. Explicit descriptors, a clearly named role, and a specific task structure are what separate a useful agent from one that needs constant correction.

The prompt builds the playbook. The playbook briefs the agent. The agent does the scanning while you focus on the work only you can do.

---

What to try next

1. In the next 30 days — load your context and run the full CRIT prompt. Don't describe your role. Paste your actual job description, your team size, your industry, and your company context. Let AI ask its three questions. Answer them slowly — especially the ones you didn't expect. The playbook you get back is only as precise as what you put in.
2. In the next 60 days — define the ROIs your business actually cares about and map your stakeholders. The base prompt doesn't automatically generate financial framing. Add an explicit instruction: "Factor in the objections I'm likely to receive and the ROI indicators my business cares about." Whether you're making an internal case or an external one, the business outcome language is what gets you in the room — and keeps you there.
3. In the next 90 days — brief a personal agent on your highest-value descriptors. Take the playbook output and turn it into agent instructions. Be explicit about the type of work, organizations, and opportunities you want it to track. Start with one agent. Test it. A well-briefed agent running in the background frees you to focus on the diagnostic and design work that no prompt can do for you.