"We frame it to an audition, not an interview." — Marty Murrillo, Precisely.
That line from Marty cracked the room open. The business problem on the table at our May 14 AI Practical Lab wasn't tools — it was identity. AI is shifting where talent professionals create value, and hiring managers are increasingly putting candidates in front of a screen and saying show me how you use AI to solve this. Waiting for a corporate framework to catch up isn't a strategy. Each leader has to actively rebrand themselves as someone who drives business impact in an AI-enabled workplace — before the market decides for them.
Who was in the room: Nicole DeFalco (Upsurge Advisors), serving as the Internal Client who grounded the situation in what the business actually wants; Marty Murrillo (Precisely), our AI Strategist and AI Learning Guru walking the room through live prompt-building; Patti Ouzounian (ELE), co-facilitating and naming the design lens; Dirk Tussing (ELE) anchoring the broader frame; plus senior talent leaders, L&D heads, organizational effectiveness practitioners, and consultants from across the ELE community.
This was a hands-on lab — not a talk. Everyone built their own prompt in real time using the CRIT model (Context, Role, Interview, Task) from Geoff Woods' The AI-Driven Leader, fed it into their LLM of choice, and watched AI interview them before generating a personal playbook. The output was uncomfortable for several people in the room — and that turned out to be the point.
The CRIT Model for AI-driven leadership, from Geoff Woods.
The interview step is the lever — and the one most people skip
When Marty walked the group through CRIT, he kept coming back to the same diagnosis: most leaders don't get value from AI because they ask it to perform before they let it understand. Context first, then questions, then output.
The Interview step — telling the AI to ask you questions before producing anything — is what shifts AI from a glorified search engine into something closer to a thinking partner. Dirk named the shift out loud while watching it work in his own chat window:
"We can use ChatGPT like a glorified Google, or we can use this as a thought partner." — Dirk Tussing
The practical move for senior leaders isn't "use AI more." It's structural: stop treating AI as a generator and start treating it as a coach that surfaces your blind spots before you commit to an answer.
AI exposes the gap between your title and your value
When members pushed CRIT into their own contexts, several got hit hard by what AI surfaced — not in a bad way, but in a "I can't unsee this" way. Cecilia Lillegard captured the room's quiet discomfort in the chat, and Nicole read it out:
"The discomfort I felt was how unprepared I was for the self-reflection of my impact." — Cecilia Lillegard
David Scherer named the same thing after his AI pushed back on his framing:
"These questions here are something I've never asked myself." — David Scherer
The takeaway for leaders isn't "use AI for self-reflection." It's harder than that: if your AI thought partner is asking sharper questions about your value than your manager or your last interview did, you have a positioning problem — and it's yours to solve, fast.
The new prep is repeatable — and you do it before the interview
Marty's audition reframe wasn't a metaphor. It's already showing up in real hiring loops: here's the problem, share your screen, show me how you use AI in real time. If the last time you "practiced" was rehearsing answers to behavioral questions, you're behind.
DeBorah Lenchard surfaced the principle that should anchor all of this — and the way she landed it was that she was quoting her own AI output back to the group:
"Use AI to accelerate, but you don't outsource accountability." — DeBorah Lenchard
That's the line that separates the leaders who'll come out ahead from the ones who'll get lapped. AI as the accelerant. Judgment, accountability, and business framing stay with you.
What to try next
- Build one CRIT-formatted prompt this week. Context, Role, Interview, Task — draft it in a Word doc, then paste into your LLM. Early evidence: the interview questions it asks should make you uncomfortable. If they don't, your Context block is too thin. Add more.
- Run the same prompt across two LLMs and compare. ChatGPT, Claude, Gemini, Copilot will produce noticeably different outputs from identical input. Early evidence: one will already "know you" based on past use — that tells you which one to invest in deepening.
- Turn your best prompt into a reusable agent or template. If you're rewriting the same context block more than twice, build it once as a custom GPT, an instructions file, or even a PDF you can drop into any chat. Early evidence: prep time for a strategic conversation drops from 45 minutes to under 10.
A note on what's changing at ELE
Starting in July, the calendar shifts. Instead of pre-scheduled topics, ELE members will community-source the calendar by submitting real business problems they need solved. Submitters become the Internal Client. We recruit Strategists and Skeptics from the community. Sessions get recorded — and you walk away with portfolio-grade footage of yourself working a real problem with peers.
The May 14 lab was a preview of that model. The room voted with their attention.
If any of these insights resonate—and you've got a top-of-mind talent business problem you'd like the ELE community to work on—send it our way.
Members can submit their challenge here: Submit My Challenge Now
Looking to turn these ideas into action? Access the Idea Exchange Post-chat for an updated Action Planning Worksheet.
