Who talked, who asked, and what got promised.

Talk share per person, questions asked against questions fielded, interruptions, the topics that kept coming back, and the commitments somebody made out loud. All of it counted from the words rather than guessed by a model.

Cedar Build, handover review48 minutes, four people, 103 turnsNobody took a noteOpen
Talk share44 percentone voice
Questions18 per 1kasked, not just said
Commitments11 per 1ksaid out loud
Interruptions12half from one person
Read off the words, not guessed.
After the meetingWho talked, who asked, and what got promised

It writes down every turn, then reads the transcript back: talk share per person, questions asked against questions answered, and the commitments somebody made out loud.

The read

Four ways to read a meeting.

Who did the talking, how the meeting moved from opening to close, who interrupted whom, and the words that kept coming up.

NearSyncWorkspaceMeeting IntelligenceSearch, plan and askKAll hubs
Cedar Build, handover review48 minutes · 4 people · 103 turns
No AI used
Questions18per 1,000 words
Commitments11per 1,000 words
Figures quoted7per 1,000 words
Overall toneDecisivefrom the words used most
PersonTurnsQuestionsLongest runShare of the talking
Dana4114190 words44%
Rami336240 words33%
Priya229110 words18%
Omar7180 words5%

It works on every meeting you have ever recorded.

Why no model

Free to run, on as many meetings as you like.

Because no model is involved, there is no per-meeting cost and no credit cap to watch.

  • Read six months of old recordings tonight
  • Pay nothing per meeting, and watch no credit meter
  • Get the same numbers next quarter as you got today
  • Argue from a figure, because it cannot have been invented
  • Trust it on messy input, because it was built against real transcripts
The analysisNo AI
Model usedNone
Runs onEvery past meeting
CostsNothing
Same answer twiceAlways
Your back catalogue
Recorded before todayStill readable
Re-runAs often as you like
Nothing inventedcounted from the words themselves
What it finds

Talk share and questions, side by side.

Two separate counts, so you can see whether the person doing the talking is also the one moving things along.

  • See talk share, turns and longest run per person
  • Compare questions asked against questions fielded
  • Watch the room move across five slices of the meeting
  • Count interruptions per person rather than sensing them
  • Spot a topic that came back four times and was never settled
Four people103 turns
Talk share44 percent, one voice
Questions asked18 per 1,000 words
Longest run240 words
Who fielded themAlso counted
Across the meeting
Five slicesOpening to close
Who carriedEach one
Counted, not judgedthe reading is yours to do
No AI

The analysis runs without AI.

Every number is counted from the words themselves. That is a smaller-sounding claim than a clever one, and a much more useful one: it costs nothing, it gives the same answer every time, and it cannot make anything up.

The analysis
Model usedNone
Cost to runNothing
Same answer twiceAlways
Counted, not inferred
Your back catalogue

It reads every meeting you have already recorded.

Anything that needs a model has to run per meeting, so your old recordings stay a pile nobody opens. This runs on the transcript you already have, so tonight you can read six months of meetings you had before you bought anything.

Retroactive
Recorded before todayStill readable
Re-runAs often as you like
Per-meeting chargeNone
The pile becomes a record
What it finds

Who talked, who asked, and who kept getting cut off.

Talk share per person, questions asked against questions fielded, how the room moved from opening to close, interruptions per person, and the topics that came back four separate times because they were never actually settled.

One meeting
Talk share44 percent, one voice
Interruptions12, half from one person
Came back 4 timesHandover
Listening, as a number rather than an impression

Access

Who can do what.

Anyone in the meeting can read the analysis of it. It shows nothing that was not said in front of everybody.

See

The analysis of meetings you were in.

Do

Re-run it over your own past recordings.

Manage

Read across the meetings your team ran.

Admin

Decide how long transcripts are kept.

See Meeting Intelligence on your own data.

Half an hour, and bring a recording of a real meeting. The analysis runs on it there and then.