The summary appears in the Summaries sub-tab, and as an item in your Cockpit under Ready to read.
What a summary contains
The write-up. What happened, in prose, for somebody who was not there.
Key decisions. Pulled out separately, because a decision inside a paragraph is a decision somebody will dispute later in perfectly good faith.
Action items. What people said they would do, with the owner and the deadline where those were actually said.
Attendees, each marked internal, external or guest, with their share of the talking.
The scorecard, where the meeting had a type with outcomes, showing each one measured.
Tone and topics, which are more useful across many meetings than within one.
Reading it well
Start with the decisions. They are the shortest section and the one most likely to differ from your memory of the conversation.
Then the action items, and act on them now. This is the only part with a deadline attached, and it is the part people skip.
Then the scorecard, if there is one. Not to grade the meeting: to notice what did not come up. An empty budget outcome on a discovery call is the genuinely useful signal, and it is exactly the sort of thing nobody notices in the moment because the conversation flowed perfectly well without it.
Skim the write-up. If you were there you probably do not need it. If you were not, read it first and skip everything else on a busy day.
Turning promises into tasks
Every action item can become a real task in one click: an owner, a due date, a place in somebody's Plan.
Do this in the same sitting. An action item is a claim that somebody said something. A task is work that exists in a system. The gap between the two is where the large majority of meeting follow-through dies, and it is a gap of roughly four seconds.
The task keeps its link back to the meeting, so months later you can answer where it came from and, if it matters, read the sentence that created it.
Not every action item deserves a task. Some were politeness. Some were somebody thinking out loud. Some are already done by the time you read the summary. Promote what is real and leave the rest, or your Plan fills with things nobody ever intended.
The talk share
Attendees show what share of the conversation each one carried.
Read crudely this is a vanity metric and worth ignoring. Read properly it is a diagnostic, and the interesting cases are the extremes.
A discovery call where you did seventy per cent of the talking did not discover very much. A review where one person said almost nothing did not review anything. Neither of those is obvious from inside the conversation and both are obvious here.
The deeper version of this analysis lives in the Intelligence sub-tab.
What participants get
Where your organisation sends a summary outward, external participants receive only the account of the meeting: what was discussed, what was decided, what was agreed.
They do not receive outcome scores, internal analysis, or coaching. That separation is enforced in the pipeline rather than left to whoever is pressing send at five in the afternoon.
The practical consequence is that you can measure honestly. If outcome scores could reach a customer, no team would score honestly, and the whole exercise would collapse into flattery within a quarter.
Finding old ones
Summaries are searchable, and the assistant can be asked about them directly. What a company said about pricing. When something was last discussed. What was agreed with a named person and on which date.
That is the compounding value of the whole system. Any single summary is worth about ten minutes. Two hundred of them is an organisational memory that does not depend on who happened to be in the room, or on whether that person still works here.
Where analysis stops
The write-up reflects what was said, including where what was said was wrong. It does not check claims against your data, and it does not know that the figure somebody quoted from memory was six weeks out of date.
Treat it as a faithful record of a conversation, which is exactly what it is, rather than as a verified account of the world.
Sentiment and topics
Alongside the write-up, each meeting carries a read on its tone and the topics that dominated it.
Both are close to useless for a single meeting and genuinely useful across many. One tense conversation means somebody was having a bad afternoon. A run of tense conversations with the same account is a signal, and it is one nobody notices from inside the sequence because each individual meeting had a reason.
Treat them the way you would treat any aggregate: ignore the point, watch the pattern.
When the summary is wrong
It happens, and there are two different kinds of wrong.
It missed something. Usually because the thing was implied rather than said, or said by somebody the transcription attributed badly. Add it yourself; the summary is a record you can correct.
It recorded something inaccurate. Almost always because somebody said something inaccurate. The summary is faithful to the conversation, not to reality, and the fix belongs in the record the claim was about rather than in the write-up of the meeting.
Knowing which of the two you are looking at saves a lot of pointless irritation with the tool.
Who should read which part
If you were in the meeting, read the decisions and the action items, and skip the write-up.
If you missed it, read the write-up first and everything else second.
If you manage the person who ran it, the scorecard is the part with information in it, and the write-up is context.
If you are picking up the account cold, read the last three summaries in order. It is twenty minutes and it is better than any briefing anybody will give you verbally.
Three habits
Read the summary the same day. The corrections you would make are available for about twenty-four hours and then they are gone.
Promote actions before you close the tab. It never happens later.
Forward it to the person who missed the meeting, rather than summarising it verbally. It is better than your summary and it takes no time.
Did this answer your question?
No, ask a person