AI Email Analysis
A thread is only read when somebody on it is already a record of yours and has actually written something. No match means the model is never called at all, rather than called and then discarded.
Which threads get read
analysedThe whole thread is read once and the result is keyed to the thread, so re-analysing updates the same row rather than piling up.
- summarywhat the thread is about and what was promised
- sentimentpositive, neutral or negative
- intentionswhat the contact wants, what you committed to
- spam_urgencyhow likely a blast, and how fast a reply is needed
- tonalitya call concept, not run on mail
- scorescoaching scores belong to calls
Product figures from the platform’s own defaults - not customer averages
How it works.
The contact gate is hard, and it runs on the server
Every sender, recipient and copied address is lowercased and looked up against the primary and secondary mail on your leads. Without a hit the function returns skipped and no request is made to any provider.
Threads that only went outward are left alone
At least one message has to have come from the other side. A sequence that has gone out and heard nothing back is not yet a conversation, so there is nothing there to read.
Four dimensions, and two left out on purpose
Summary, sentiment, intentions, and the spam and urgency reading. Manner of speech and the nought to ten coaching numbers are call ideas, so this path neither asks for them nor stores them.
A long thread keeps its tail
Messages render oldest first with direction, timestamp and parties, then the whole block is trimmed from the front if it runs long. The thread identifier hashes to a stable value, so reading it again updates one row.
Two moments in every read.
Every read passes through the same seven. AI Email Analysis 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- Gate
- One participant address must match a lead. No match means no request at all
- Also required
- At least one message received on the thread. Drafts are excluded from the read
- Dimensions
- Summary, sentiment, intentions, spam and urgency. No manner, no coaching numbers
- Input
- Up to twenty four thousand characters, keeping the most recent messages
- Trigger
- A message received, once per thread per quarter of an hour. By hand it skips that wait
- Identity
- A hash of the thread, so reading it again replaces the row rather than adding one
- Which model reads it
- The mail purpose resolves through the model catalogue like every other read, so it is one decision made once rather than per thread
- Not the same as
- This reads mail you already own. Composing a reply is its own capability
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.
The model proposes record updates with its confidence and the quote it heard them in, shown as a before-and-after you accept or reject row by row.
One to six sentences on the record covering why the call happened and what was agreed, so nobody reading the list has to open the audio to find out.
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.
Callbacks and meetings pulled out with their times, a confidence figure and the words they came from, ready to accept in one click or reject in one.
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.
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 ai email analysis on your own floor.
Thirty minutes, your numbers and your data. We will set ai email analysis up live and you can decide from the thing itself rather than from this page.
14-day trial · no card · migration included