Check AI 1:1 prep against your notes

Before you use an AI suggestion in your next 1:1, find the part of the written record that supports it and check what changed afterward. Then ask whether it improves the question you would have prepared yourself. A fluent suggestion can still send you back to a problem already resolved.

1on1 uses written session context, prior follow-up and the manager-only addendum to help prepare the next conversation. That gives you a specific job to evaluate: does a suggested next step fit this pair’s current situation?

Prepare once without AI

In a public discussion about 1:1 preparation, a manager described using Excel prompts and rereading the previous notes, then asked what else people did. That is a useful starting point for your own comparison.

Before opening the suggestions, reread the relevant sessions and actions. Write down the question you would bring, the note behind it and anything that has since changed. Keep that small baseline: otherwise, an impressive-looking list can replace your judgement before you have made it.

Use only a workflow your organisation permits for these records. For an initial demonstration, use fictional material rather than uploading employee notes to a public AI service.

Test what changed

Here is one fictional practice sequence you can adapt.

First session: An engineer is waiting for a test environment. The manager agrees to request access. The notes also say the engineer wants to lead a migration.
Next session: Access has been granted and that action is closed. A separate dependency on another team remains unresolved. The engineer corrects the earlier note: they want to contribute to the migration, not lead it.

Your manual preparation should now focus on the remaining dependency. “Which decision are we waiting for, and who can I ask?” is a possible question to bring.

“Have you requested environment access?” repeats closed work. “Are you ready to lead the migration?” carries forward the corrected assumption. Reject those premises, even if the wording sounds thoughtful. The useful distinction is between something mentioned before and something that still needs attention.

Keep a small decision log

Now inspect the actual suggestions from the approved workflow. For each one you consider using, record:

Evidence: Which session or action supports the premise? Check the latest relevant update too.
Decision: Keep, reject or rewrite, with a short reason.
Your question: What will you actually ask, in your own words?

You can keep this log beside your preparation notes. If you cannot establish a suggestion’s premise, leave it out or ask an open question that lets the report clarify it.

Compare the result with your manual baseline. Did AI surface a relevant follow-up you missed, or merely rephrase it? Did checking stale suggestions create more work than it added? Preparing manually remains a reasonable choice. You are evaluating the assistance, not scoring the person.

Ask 1on1 to show the preparation workflow

1on1 starts with a 7-day full-feature trial.

Ask 1on1 about a demonstration: “We want to understand how preparation handles an unresolved dependency, a closed action and a corrected assumption. Can you show the supported workflow with fictional records and explain current AI access?”

Bring your manual baseline to that demonstration. Keep the suggestions that earn a place in the next conversation.