AI capability

Questions in English, answers with receipts

The value is not that you can type a question. It is that the answer arrives with the transactions behind it, computed within your permissions, and that the system declines when the data does not support the question rather than producing something plausible.

Rows
DeliverySalesG&A
New York$412K$286K$104K
Austin$238K$141K$62K
Remote$176K$88K$39K
Every dimension is on the journal line, so any combination is a query rather than a rebuild.
Answers cite transactionsScoped to the askerDeclines rather than approximates

What it does

Six things, specifically.

Understand the question

Which vendors did we pay more than $50,000 to last year. What is our largest unbilled project. How many customers pay later than their terms.

Query the graph, not a report

It composes against the underlying model rather than picking from a catalogue of pre-built reports, so the question does not have to have been anticipated.

Return the evidence

Every figure links to the transactions that produced it. A number you cannot open is a number you have to trust, which is the wrong relationship.

Respect scope

Computed within the asker’s permissions in the data layer — not computed broadly and then redacted, which leaks the aggregate.

Become a report

Any answer saves as a view, schedules as a report, or exports. Most people use conversation to discover the report they wanted.

Refuse cleanly

Unclosed period, missing dimension, ambiguous phrasing — each produces a specific explanation rather than a confident approximation.

Why the receipts matter more than the language

Natural-language querying has been demonstrable for a while and is not, by itself, the interesting part. The interesting part is whether you can check the answer without rebuilding it.

A number returned without provenance has to be trusted, and finance people are correctly unwilling to trust numbers they did not derive. So every answer here carries the filters applied and the transactions included, one click away. The first thing most new users do is open the transactions to see whether the figure holds — and after a fortnight of it holding, they stop.

The test of a query interface is not whether it understood the question. It is whether you can check the answer in ten seconds.

Ambiguity should produce a question, not a guess

Ask about revenue last quarter and there are at least three defensible readings: booked, recognised, or collected. A system that silently picks one is wrong a third of the time in a way nobody notices.

Ours asks which you meant, once, and remembers the preference. That is marginally slower and considerably more useful, and it is the same principle as the copilot declining rather than approximating.

Where the permission boundary sits

Scope is applied when the query is composed, not to the results afterwards. The difference matters: a system that computes a company-wide total and then hides the departmental breakdown has already exposed the aggregate, which is frequently the sensitive figure.

A department head asking a company-wide question gets a clear refusal naming the scope limit, rather than a partial answer they might mistake for a complete one.

Limits

Where it does not help.

Every capability page on this site carries one of these, because a feature described without its boundaries is a claim rather than a description.

It is not a forecast

Questions about the future route to explicit modelling in the CFO agent. Conversational extrapolation is where confident nonsense comes from.

It cannot answer what was never captured

No phrasing recovers a department that was never coded. It explains that rather than deriving something adjacent and presenting it identically.

It is read-only

It cannot post, approve, or change anything, at any authority level. Asking returns a refusal and a link to the workflow that does it under controls.

Questions

What people ask.

How is this different from the copilot?
The copilot is the conversational surface; this is the query capability underneath it, also available through search and the API. Practically, most people meet it as the copilot.
What if it misunderstands the question?
It states the interpretation it used alongside the answer, so a misreading is visible immediately rather than discovered later. Where two readings are equally defensible it asks.
Can it query across entities?
Yes, within the asker’s scope, with entity-level breakdown available on any consolidated figure.
Does it work on our current ledger?
Yes. It queries the business graph, which reads your existing accounting system and surrounding tools.
Can we build reports from answers?
Yes — save, schedule, or export. Conversation is a good way to find the report you wanted and a poor substitute for having it.

Ask it something you actually need.

Two periods of data and one real question is enough to judge whether the answers hold up.