Follow-Up Extraction
A promise made out loud is worth nothing until it sits on a day. Callbacks and meetings are pulled off the call with whatever time was genuinely said, and become real work the moment somebody accepts one.
Follow-ups pulled off the call
1 to review- callbackCall Dana about the renewal88%Thu 09:30accepted
can somebody call me Thursday about the renewal
- meetingRenewal review, CFO joining72%Thu 14:00 to 14:45
book me Thursday and I will bring our CFO
- still listeningcallbackConfirm the credit has landed55%no time stated on the call
let me know once the credit is through
- 09:00Callback09:30
- 10:00
- 11:00
- 12:00
- 13:00
- 14:00
- 15:00
Product figures from the platform’s own defaults - not customer averages
How it works.
Two kinds, and nothing beyond them
A callback or a meeting. Which one it is decides the task type on the other side, so a meeting arrives as a meeting task rather than as one more number to ring.
Times are optional and are allowed to stay missing
Due, start and end may all be absent. Where nobody named a day, the row arrives without one instead of with an invented Thursday, because a diary full of guesses is worse than a diary with holes in it.
Accepting creates the task and gives it an owner
Title, type, dates and notes are written, assigned to whoever took the call, or to the person reviewing when the call had no agent. An assignee record follows so the work lands in a genuine queue rather than only in an administrator's list.
Repeats are caught on the title
Before a row is written the call is checked for a pending or accepted follow-up of the same kind and title. Analysing the same conversation a second time therefore does not book anybody twice.
Two moments in every read.
Every read passes through the same seven. Follow-Up Extraction is the lit ones, and everything either side of it is a different page in this category.
- 01Source
the call or the thread it reads
- 02Transcribe
audio into words, with speakers
- 03Read
the pass over the whole of it
- 04Judge
the score, the sentiment, the intent
- 05Extract
the fields and follow-ups pulled out
- 06Write
what lands back on the record
- 07Review
a person checking the machine
The specifics.
8 facts- Kinds
- callback and meeting. A meeting may carry a start and an end, a callback a due date
- Times
- Optional throughout. A follow-up with no stated time is created without one
- Each row carries
- How sure the model was, between nought and one, plus the words it heard the promise in
- Accepting
- Writes a task owned by the call's agent, or by the reviewer where there was none
- Linking
- Attached to the lead when a number on the call matches one, and left loose otherwise
- Bulk accept
- Asks for confirmation beyond five rows and takes the low confidence ones with it
- Re-analysing
- Looks for the same kind and title on that call before adding anything
- Not the same as
- This creates work. Field suggestions edit columns on the lead instead
What it reads, and what reads it.
Nothing here invents its input. These are where the material comes from, and where the verdict goes afterwards.
More in Intelligence
12 capabilitiesAI that proposes edits to the record - an insight becomes a field once you accept it.
One reading of the transcript returns the summary, the sentiment, the tone, what each side intended and a set of scores, in a shape fixed in code.
Positive, neutral or negative on every analysed call, stored in a column of its own so the ones that went wrong are a filter rather than a listening exercise.
Chosen words rather than a fixed list for how the agent sounded, how the customer did, and the conversation overall - the manner behind the sentiment.
Communication, resolution, satisfaction and compliance scored nought to ten beside a separately judged overall, on every call that clears the analysis gates.
Compliance scored nought to ten against what the vertical says matters, because no phrase list exists to recite. A missing reading is excluded, never passed.
Four of six speech paths return the channel each side was recorded on, so the transcript knows who spoke rather than guessing, and one control swaps a pair.
Collections, healthcare, insurance, admissions and eleven more - each shaping what the analysis looks for without letting anyone break the output shape.
Threads summarised and read for intent and urgency, but only when a participant matches a record, so the model is never called on mail that is not yours.
Draft from an instruction or rework a message you have, landing in the composer with a suggested subject and going nowhere at all until you send it.
Describe what you track or paste a spreadsheet's columns, and get a named field group back with types, options and placeholders to approve one at a time.
A sentence becomes a grid filter and a sort order with the reading explained back, on the toolbar of every grid that carries a column menu of its own.
OpenAI, Azure, Anthropic, Gemini, Bedrock and ten more, bound per purpose with an ordered fallback chain, a circuit breaker and a cost line on every attempt.
The rest of the platform.
Five more categories, all on the same record and the same bill. Each card names three of its capabilities, so you can tell from here whether it is worth opening.
Intelligence
See follow-up extraction on your own floor.
Thirty minutes, your numbers and your data. We will set follow-up extraction up live and you can decide from the thing itself rather than from this page.
14-day trial · no card · migration included