How AI-fluent is your leadership team?

April 23, 2026

---

Hi Everyone,

You might be comfortable with AI yourself. When was the last time you watched your CFO or head of operations actually use it?

Russell Reynolds surveyed over 3,000 leaders in H1 2025. 82% said AI skills are essential for C-suite success. Only 41% felt confident they could put AI to work. The expensive mistakes happen when the people approving AI spend haven't used the tools enough to know what they're approving.

Training alone doesn't fix this. A 2026 Aalto University study found that AI training made leaders more confident but less accurate about what they actually knew.

Here's a fluency scale for your leadership team, and two things to do about it this week.

A fluency scale for your leadership team

Ask each person on your leadership team to pick the level that honestly describes them today. Include yourself.

Tier 1 – Awareness: Has tried an AI tool a few times. Can define "hallucination" but wouldn't catch one in a financial summary or a customer email. Leaders at this level tend to rely on vendor demos to decide what AI can and can't do.

Tier 2 – Working knowledge: Uses AI at least five hours a week on actual work. Can write a good prompt, spot mistakes in AI output, and name the specific tasks where AI saves them real time. Without that firsthand experience, it's hard to tell whether a new tool is worth the investment. VPs and Directors need to be here because they're the ones deciding how their teams use these tools day to day.

Tier 3 – Strategic fluency: Can decide what to delegate to AI across a business function, set the team's AI usage policy, evaluate vendors, and brief the board without help from the technical team.

You can test yourself here: could you walk your board through your AI spending tomorrow without needing your technical team in the room?

Ethan Mollick at Wharton sets a floor of 10 hours a week of hands-on use for leaders at this level.

Tier 4 – Technical fluency: Reads model benchmarks, understands how retrieval and generation systems work, and participates in build-versus-buy decisions at the architectural level. Your CTO or Chief AI Officer needs this, but it's optional for everyone else.

Find your hidden AI users

WalkMe and SAP found in 2025 that 49% of employees hide their AI use from management. Another 45% pretend to understand AI tools in meetings when they don't.

Run a lunch and learn session where everyone demos how they're actually using AI. Make it clear up front that nobody will get in trouble for using tools that weren't officially approved. The point is to learn what's already happening, not to police it.

Mollick calls these people "secret cyborgs," and says their prompts and workarounds are usually better than anything you'll get from a training vendor. Right now, that knowledge is invisible to you.

Build one shared framework (with this AI prompt)

KPMG uses three words across their entire firm – Think, Prompt, Check. Think about whether this task is right for AI, write a clear prompt, then check the output before you use it. HubSpot built an AI fluency rubric into hiring and quarterly reviews, and 85% of employees now use AI monthly. Zapier took a similar approach for 800 employees and reached 100% adoption.

We built an AI prompt that helps you create your own version. You paste it into Claude, ChatGPT, or Gemini, answer a few questions about your company, and it produces a one-page framework with custom fluency levels, a plan for where to embed it, and a name your team will actually remember. The whole thing takes about 15 minutes. Download it here.

Build your AI Fluency Framework

Go deeper

👉 Ethan Mollick: Making AI Work – Leadership, Lab, and Crowd – the clearest framework for turning scattered AI experiments into real adoption across an organization

👉 Russell Reynolds: Leaders' Views on Generative AI in 2025 – the full survey behind the 82/41 split, with breakdowns by role and region

👉 Anthropic: AI Fluency Framework & Foundations – a free course built around four fluency competencies (Delegation, Description, Discernment, Diligence)

👉 McKinsey: Superagency in the Workplace – where the C-suite blind spot data comes from, and what high-performing organizations do differently

Coming up tomorrow

Tomorrow, we're covering the five things you need to write down when your team starts working across borders.

P.S. Forward this to your leadership team and ask each person to reply with their tier number.