An AI that actually reads the standups
Nobody reads twelve answers a day, which is how a written standup quietly becomes write-only. The AI reads them and tells you what moved, who is stuck and what needs attention before it bites.
Summary
Billing webhook migration landed and the search relevance fix is out. The index rebuild starts today.
Blocked: Tomas on the Stripe key rotation, second day running. Needs someone with production access.
Worth a look: the same deploy blocker has now come up in three of the last five runs.
9 of 10 answered · full answers below
Three things a summary has to do to be worth reading
A digest that just concatenates everyone's answers is the thing people already stopped reading.
What actually shipped
Pulled out of the answers as concrete work, not restated as a bullet list of everyone's paragraph. Twelve updates become the four things that moved.
Who is blocked, and since when
Open blockers are tracked across runs, so a blocker in its fourth day reads differently from one raised this morning.
Patterns worth a look
Response-rate dips, recurring blockers and overtime patterns surfaced while they are still small enough to fix.
Blockers get tracked, not just mentioned
A blocker written down on Monday and again on Thursday is one problem, not two updates. The digest carries open blockers between runs, so the age of a blocker is visible without anyone scrolling back through a week of channel messages to work it out.
That is the difference between a standup you read and a standup you act on: nobody has to notice the repetition themselves.
The three ways teams do this in Slack, with and without a bot, are compared in How to track blockers in Slack.
The summary outlives the channel
Connect Notion in Settings and every closed run is archived as a page in a check-ins database asyncmate creates for you: the AI summary on top, the blockers and what shipped, then everyone's answers and who did not respond. A quarter of standups becomes a database you can filter by check-in, date or blocker count, instead of a channel you scroll.
How to connect it and what lands there, in Send check-ins to Notion.
Included in Pro, with no separate AI bill
The AI digest is the reason Pro exists, so it is not metered on top of it. You pay $2.50 a month per person who answers a check-in, and the digest is part of that. There is no API key to bring, no token counter, and no per-summary charge.
A per-workspace daily allowance scales with team size and keeps a runaway install from becoming a surprise. Past it, digests keep posting in the plain format.
See what a team your size would pay or read what counts as an active participant.
AI standup digest FAQ
- What is an AI standup digest?
- The summary at the top of the message asyncmate posts when a check-in's response window closes. Instead of reading twelve people's answers in full, you read a few lines covering what shipped, what is blocked and what needs a decision, with everyone's raw answers still underneath.
- Which AI model writes it?
- A configurable model, Claude Haiku by default, chosen because a digest costs cents to write at that tier. We hold the API key, so there is nothing for you to set up or pay for separately.
- Does the AI digest cost extra?
- No. It is included in Pro at $2.50 per active participant per month, with no separate AI charge and no token metering. The free plan posts the same digest without the summary on top.
- What language does the summary come out in?
- The one your team answers in. If people write their updates in Croatian, the summary is in Croatian; the labels around it follow too. Nobody has to switch to English so the bot can keep up.
- Is there a limit on how many AI digests we get?
- Each workspace has a daily allowance that scales with team size, roughly one AI digest per day for every five active participants. Past it, digests still post in the plain format rather than failing. You can also switch AI off for a specific check-in and leave it on everywhere else.
- Do you train models on our standup answers?
- No. Answers are sent to the model to produce your digest and nothing else. We do not train on them and we do not use them to improve any model. What is sent is the run being summarized plus the team's still-open blockers, which is how the digest can say how long something has been stuck.
- Does the AI summary go into Notion too?
- Yes. With Notion connected and Send to Notion switched on for a check-in, every closed run becomes a page in a check-ins database asyncmate creates in your Notion, with the AI summary at the top and everyone's answers under it. The summary is written once, when the digest posts, and archived as is.
- Can I turn the AI off?
- Yes, per check-in and workspace-wide. Turned off, the digest posts exactly as it does on the free plan: everyone's answers, no summary, same channel, same schedule.
Stop reading standups, start reading the summary
Standups run free for up to 10 people who answer. The AI digest comes with Pro, and you can switch it off per check-in whenever it is not earning its place.
The rest of the same install
- How the Slack standup bot works
- Logging hours by project in Slack
- Running a one-tap poll in Slack
- How anonymous surveys stay anonymous
- Giving kudos in Slack
All of it is one app, in Slack or Microsoft Teams. See every feature or check what a team your size would pay.