Databricks AI integration: lakehouse data in your finance workflows

Finance teams next to a lakehouse all know the queue. The data is right there, every order, every transaction, every usage record, but getting an answer out of it means filing a ticket and waiting for someone on the data team to find time between their own priorities. So the monthly check against the books slips, the export from last quarter gets reused, and the question "are we invoicing everything" gets answered with a shrug and a sample.

Nance gives finance a direct line. It queries Databricks through a service principal you control, checks operational datasets against the administration, verifies revenue completeness against real order and usage data, and answers ad hoc questions from the tables you exposed. The data team keeps owning the platform; finance stops borrowing their calendar to use it.

What Nance automates with Databricks

Operational data versus the administration

Nance checks what your lakehouse datasets recorded against what was booked, and surfaces the differences with the underlying rows attached.

Revenue completeness checks

Order, usage and transaction data becomes the reference for what should be in the books. Missing invoices show up as findings, not as year-end surprises.

Source detail for bookings

When a booking needs operational backup, Nance pulls it from the relevant tables directly instead of routing the question through another team.

Reporting on governed data

Finance reporting draws on the same curated Databricks data the rest of the company uses, so the monthly report and the company dashboard tell one story.

Scheduled consistency checks

Recurring comparisons between lakehouse and ledger run on a schedule, surfacing drift during the month instead of during the close.

Ad hoc questions, answered

The questions finance used to file as data requests, Nance answers directly from the datasets you exposed, with every query logged and reviewable.

Connect Databricks to Exact Online, Twinfield, Moneybird & Xero

Most Databricks connectors are data pipes: they copy fields on a schedule and stop there. Nance is the connection and the colleague in one. It reads Databricks, applies your rules and does the actual finance work in your accounting package, with approvals where you want them and every action logged.

Frequently asked questions

How does Nance connect to our Databricks workspace?

Through a scoped, read-only connection by default: a dedicated service principal you create and control, limited to the catalogs and schemas you choose. Write access only where you explicitly grant it.

Does Nance build pipelines or manage our lakehouse?

No. Nance queries the data your team already maintains. It reads, reconciles and reports; data engineering and platform management stay with your data team.

What should we expose to Nance?

Most teams grant access to a small set of curated tables or views relevant to finance: orders, transactions, usage. Narrow access keeps queries simple and answers easy to trust.

Can we audit what Nance does in Databricks?

Yes. Every query is logged with its context and result, and access is bounded by the permissions of the service principal you set up. Your own workspace audit logs see Nance like any other principal.

Finance answers from the lakehouse, without the ticket