An AI tool for 1-on-1 meetings that builds the record for you

Jot a quick note after each 1:1, and AI turns it into the wins, the dated commitments, and the topics your report raised — proposals you approve. Over a quarter it becomes a living record of every direct report, instead of a folder of notes you never reopen.

Most managers leave a 1:1 with good intentions and a fuzzy memory. Someone mentioned they're stretched thin; you meant to follow up. Someone agreed to ship the handoff by Friday; you can't quite remember if it was this Friday or last. A month later the details are gone, and at review time you're reconstructing a year from the three weeks you happen to recall. An AI tool for 1-on-1 meetings should fix exactly that gap — not by sitting in the room recording you, but by catching what mattered the moment after you've said it.

Capture a note, not a transcript

There's no bot in your meeting and nothing is recorded. The 1:1 stays what it should be: a private conversation between you and one person. When it's over, you paste a few lines in whatever shorthand you already think in — "Maya crushed the migration, still blocked on the design review, wants to lead the next launch, I owe her a headcount answer by Fri."

That's the whole input. You don't fill in fields or tag anything. The AI reads the note the way an attentive chief of staff would and hands back structured proposals.

AI proposes wins, commitments, and topics

From that one paste, the tool pulls out three kinds of things and presents each as a proposal you can accept or wave off:

• Wins — the accomplishments worth remembering. "Led the database migration to completion." These are the raw material your future performance review is made of.
• Commitments — the promises, both directions, with a rough date attached. "Manager to deliver headcount answer by Friday." Nothing erodes trust faster than forgetting what someone told you; nothing builds it faster than reliably following through.
• Topics — the things your report raised: a blocker, an ambition, a friction with another team. "Wants to lead the next launch." These are the threads you reopen next time so the conversation has continuity.

Each proposal is evidence-based — drawn from what you actually wrote, never invented. You read down the list and approve the ones that belong.

Nothing is tracked until you approve it

This is the line that matters, so it's worth being plain about: the AI never auto-applies anything. Every proposal waits for a human. You approve a win, and it joins the person's timeline. You dismiss a topic that doesn't belong, and it's gone. The record only ever contains things a manager decided were true. That keeps the system trustworthy enough to base a real review on — and keeps you, not a model, in charge of what gets said about your people.

The work of a 1:1 happens between 1:1s. The point of capturing it is so the next conversation can pick up exactly where the last one left off.

A living record every other surface draws on

Because each approved item lands on that person's record, the value compounds quietly. Before the next 1:1, the commitments you both made are right there — you open with "Last time you said the design handoff was rough; how did it land?" and your report hears that their words stuck. Across a quarter the wins and trajectories accumulate, so when it's time to write a review there's no blank page and no recency bias: it drafts from evidence you already approved. The same per-person record is what the rest of Performle reads from — including the goal suggestions that propose each person's next step.

How it works

1. Finish the 1:1. No prep, no bot, no new ritual — just have the conversation you'd have anyway.
2. Paste a quick note. A few lines in your own words, however you already jot things down.
3. Approve the proposals. AI suggests the wins, the dated commitments, and the topics raised; you accept or dismiss each.
4. Pick up next time. Commitments resurface, trajectories build, and the review later writes itself from what you approved.