/11

Leader Meeting Kit

A no-keynote format for turning a room of leaders into a useful AI peer group: who to invite, what to compare, what to share, and what to test next.

Leader

6 min

Open +

The useful AI conversation is not "what is the future?" It is "what survived contact with the business?"

Leave knowing / Run a useful AI leadership session

The agenda

Sixty minutes. No keynote.

0-5 / Fear dump

What worries people about AI right now? Put it on the table so it does not quietly control the rest of the meeting.

5-15 / What is real

Each person shares one AI use that is genuinely in production or genuinely failed. No vendor demos.

15-30 / Map the work

Choose one workflow everyone recognizes. Draw the steps, handoffs, delays, approvals, and information sources.

30-45 / Find one AI job

Pick the smallest step AI could improve. Decide what stays human and what evidence would prove value.

45-55 / Compare notes

What has somebody else already learned about tools, data, governance, adoption, or failure that the group can reuse?

55-60 / Commit

One owner. One pilot. One metric. One date to compare what happened.

Who belongs

Small enough to be honest. Different enough to be useful.

Operators

People close enough to real workflows to know where the friction, exceptions, and workarounds actually live.

Decision makers

Leaders who can approve a pilot, remove a blocker, fund a useful system, or stop a bad one.

Risk voices

Someone who can surface security, privacy, legal, people, or customer consequences before they become cleanup.

Different stages

Mix people who are ahead, behind, and in the messy middle. A peer group made only of AI experts turns into a conference panel.

The community contract

Five rules that make the room worth returning to.

No vendor pitches. If a tool matters, explain the workflow and result before the logo.

Bring one real artifact: a prompt, workflow, policy, failure, metric, screenshot, or test result.

Separate what you know from what you think. Label demos, pilots, production use, and hearsay.

Do not share customer, employee, security, or commercially sensitive data you are not allowed to share.

Leave every session with one named experiment or decision, then report what happened next time.

Questions worth asking

Twelve questions for peers who are actually doing the work.

What survived the demo?

Which use case is still being used 30 days later?

What did people reject?

Where did staff ignore or work around the AI?

Where did quality get worse?

What new review burden or error did the AI create?

What did you stop buying?

Did AI remove software, agencies, process, or only add another subscription?

What data was harder than expected?

Where did access or cleanliness block the idea?

What needed custom work?

Which part could not be solved by an off-the-shelf product?

What stayed human?

Which judgment became more valuable after automation?

What got faster?

Which cycle time changed enough to matter?

What became riskier?

Where did permissions, privacy, or hallucination show up?

Who owns AI now?

Is ownership central, distributed, or still vague?

What are you measuring?

Which metric convinced leadership this was real?

What would you not do again?

The most useful answer in the room may be the failure.

Invite the room

A simple invitation that filters out spectators.

Try the template

I am putting together a small peer session for leaders who are actively trying to use AI in real work.

No keynote. No vendor pitches. No predictions panel.

Bring one thing:
- an AI workflow that is genuinely being used,
- a pilot that failed,
- a decision you are stuck on, or
- a result you can actually measure.

We will spend 60 minutes comparing what survived contact with the business, then each leave with one next test.

If that sounds useful, join us. If you only want to talk about "the future of AI," this will probably be boring.

Notes / Subscribe

No schedule. No content calendar. A note goes out only when there is something worth understanding.