Why BusinessSensor

AI that is grounded in your own books

Because the system holds your ledger rather than reading a copy of it, the AI answers from complete, double-entry-accurate data — and every draft it produces goes back through the same audited posting path a person would use.

Why owning the ledger matters here

Most AI in this category connects to an accounting package and reads what it is given. That is a fundamentally shallower position: the answers are only as complete as the export, and nothing can be written back safely. BusinessSensor is the system of record, so the AI reads first-party data that is correct by construction, and anything it drafts is posted through the same trigger-enforced path as a human entry — never around it.

What is live today

These are shipped and in production, not roadmap items.

  • A morning briefing: deterministic rules find what is worth an owner’s attention overnight, and the AI writes one grounded sentence of advice per finding
  • An ask-anything assistant that answers free-form questions from a briefing assembled server-side out of your own figures
  • AI-assisted onboarding: business type, chart of accounts, item categories, starter inventory with a real pack hierarchy, and your country’s taxes drafted for an operator to approve rather than typed from a blank screen
  • AI content: product photographs, item descriptions, storefront copy and website page drafts, all landing unpublished for review
  • In the pharmacy edition: prescription-scan capture, interaction screening and drug-data auto-populate

The privacy boundary, stated plainly

The insight engine that runs every night never sends a monetary amount out of the system — the findings are computed by deterministic rules on your data, and only the wording of the advice is generated. The assistant that does work with real figures is off by default and switched on per tenant by an administrator. You are told which is which, because a vague answer to "what leaves my system" is not an answer.

What is deliberately not here yet

Cash-flow forecasting, a CFO copilot, treasury and investment decision support, and AI-drafted transactions are designed and scoped but not shipped. They are named here as roadmap rather than implied as capability, for the same reason every other claim on this site is evidenced: the value of saying something ships is that you can trust it when we do.

Where this applies

Take this to any demo

Ask any AI feature you are shown where its numbers come from, and what leaves your system to answer a question.

An assistant reading an export can only be as accurate and as current as that export. One that cannot tell you what data leaves the building has not thought about the question. Both answers are worth more than a demo of the chat window.

Common questions

Does my financial data get sent to an AI provider?

Not for the nightly insight engine — that runs on deterministic rules over your data and sends no monetary amount out of the system. The ask-anything assistant does work with real figures to answer questions, which is why it is off by default and has to be switched on for your tenant by an administrator.

Can the AI change my accounts?

Nothing posts without a person approving it. Where the AI drafts something — an onboarding chart of accounts, a description, a document capture — it lands as a draft for review, and when approved it posts through the same audited path as any manual entry.

Is this an add-on or included?

The AI capabilities are tiered, and the assistant that works with live figures is administered per tenant. Which tier includes what is shown in the product catalog alongside the rest of the feature comparison.

See it running on your own numbers

Book a walkthrough and we will show you the system doing the job you actually need it to do.