Ask Dibein
Ask your financial data a question in plain language.
A finance-specific assistant grounded in your live forecast, ledger, and transaction history. Every answer is a number your team can trust, with the underlying transactions cited, not a plausible-sounding guess.
of numeric answers link back to source transactions or forecast rows
We almost didn't build a chat interface. Most 'AI finance assistant' products we tried before building Dibein gave fluent, confident, occasionally wrong answers about our own money — which is worse than no answer at all.
Ask Dibein is deliberately narrow. It doesn't browse the open internet and it doesn't answer questions outside your financial data. It translates a question into a query against your forecast, ledger, and transaction history, runs that query, and answers from the result — every number in its answer links back to the rows that produced it.
That constraint costs latency: a grounded answer takes longer than a free-form one, typically two to five seconds. We think that's the correct trade for a number your team is going to act on.
Ask in plain language
"Can we afford to hire two engineers in Q3?" or "Why did marketing spend jump in April?" — no query syntax or report-builder required.
Grounded retrieval, not free generation
The question is parsed into a structured query against your live forecast, ledger categories, and transaction history — the same data Cash Flow Copilot and Radar already use.
Cited answers
The response states the number and links directly to the forecast rows or transactions behind it, so anyone on the team can verify it in one click.
Escalates instead of guessing
If a question can't be answered from your connected data with confidence, Ask Dibein says so and suggests what to connect or clarify, rather than producing a plausible-sounding guess.
See Ask Dibein in action on a sample workspace.
No signup required. This runs on fixed sample data — nothing here touches your own accounts.
Natural-language forecast queries
"What's our lowest projected cash week in the next quarter?" returns the specific week, the amount, and which assumptions drive it.
Variance explanations
"Why is this month's spend higher than last month?" breaks the delta down by category and vendor, citing the specific transactions responsible.
Scenario-aware answers
Ask a what-if question and it opens a pre-filled Scenario Studio draft instead of guessing at a number, so the answer is a real model, not a hallucinated one.
Board-ready summaries
"Summarize this quarter's cash position for the board" produces a short written summary with every figure traceable to source.
Slack and email digest
Subscribe any saved question to a weekly digest delivered where your team already works.
Can it answer questions unrelated to our finances?
No, by design. Ask Dibein only answers from your connected financial data. General questions are declined rather than answered from a general-purpose model.
What happens if the data needed to answer isn't connected?
It tells you exactly what's missing — for example, a specific bank account or ledger category — instead of estimating from partial data without saying so.
Is our data used to train a shared model?
No. Each customer's data is used only to answer that customer's questions. Nothing is used to train models shared across customers.