Ask a tech firm CFO the place the quarter’s margin is touchdown, and you’ll get a straight quantity. Ask what moved it, and the checklist begins: the favored characteristic whose compute price climbed quicker than its worth, the utilization income booked throughout blended subscription and consumption plans, the reserved compute an automatic scaling coverage drew down quicker than deliberate. Each quantity on that checklist is the product of a number of techniques, and each one is more and more formed by brokers. Finance’s job is to see how these variables transfer unit economics and steer the corporate repeatedly, not as soon as a month on the shut.
Most finance groups are requested to do this on plumbing constructed for a slower enterprise: extracts, spreadsheets, and metrics that reconcile month-to-month whereas utilization, pricing, and compute transfer hourly. Every single day of lag has a worth. A repricing lands per week late. A metering bug survives till the shut. A compute dedication will get signed on final month’s image of demand.
How AI unit economics grew to become finance’s entrance line
Tech and AI-native firms are constructed on development. Utilization scales, pricing spans subscription and consumption, and compute has change into the biggest variable price within the enterprise. Defending the economics of that development has at all times been finance’s job. What modified is the velocity: brokers now form how compute is consumed, how utilization is priced, and the way income is acknowledged, and the sector’s numbers present the squeeze. AI-native gross margins reached about 52 p.c in 2026, up from 41 p.c in 2024 (ICONIQ), nonetheless in need of the 70 to 90 p.c that basic software program earns.
The strongest finance organizations are usually not ready it out. OpenAI’s finance group describes placing AI brokers to work on contract workflows, extracting and reviewing phrases that when consumed analyst hours. YipitData introduced income operations and finance onto Databricks, the place finance analysts write their very own SQL and PySpark, and a NetSuite integration took handbook effort out of economic reporting. The query for a tech firm CFO is not whether or not finance will function with AI. It’s what basis makes it protected.
Why a phrase like ontology now issues to tech firm finance
Finance has at all times been good at discovering the quantity, even when it’s buried in complexity. However a quantity will be completely correct and nonetheless not be appropriate, as a result of it rests on a partial or dated image of how the enterprise works. What issues is the which means behind it: the definitions, the merchandise and plans, the utilization and compute drivers, and the way every of these modifications because the enterprise strikes. For AI, this hole is the headline drawback: inaccuracy is the most-cited situation enterprises report with AI, named by roughly a 3rd of them (McKinsey, 2025). As Ali Ghodsi places it, most enterprise AI is guessing with false confidence. That may be a context drawback, not an intelligence drawback.
An ontology closes the hole. It captures what the numbers imply and retains that which means present because the enterprise modifications. And the uncooked materials is arriving quicker than ever. Stripe fee knowledge, transactions, subscriptions, refunds, and payouts now stream into Unity Catalog by OpenSharing on Databricks Market: stay, ruled, and with no ETL to construct or keep. Lakebase, the transactional Postgres database constructed natively on the lakehouse, extends the identical precept contained in the platform. Operational knowledge and evaluation share one basis, so the reply displays the enterprise because it runs right now, not because it was on the final sync.
The place Genie One is available in
In an AI-native enterprise, an understanding fashioned a couple of hours in the past could already be outdated. So the ontology itself has to maintain transferring. It has to be taught from the techniques the enterprise runs, sharpen with each query, and adapt because the enterprise evolves.
Databricks constructed Genie One for this. It really works as a data-smart AI coworker: a finance chief asks a direct query and will get a reliable, sourced reply, grounded in Genie One’s ontology and ruled at each step. Amagi, the AdTech platform behind 1000’s of broadcast channels, already runs this fashion in manufacturing. Its finance group will get real-time billing and monetary reporting on Databricks, with Genie One answering pure language questions, and since finance, advertising and marketing, and operations draw on the identical ruled knowledge, govt conferences not relitigate whose quantity is correct.
Each AI-native finance group is working some model of three questions, and every reply units up the subsequent:
› The place is actual gross margin touchdown, by product and buyer, as soon as AI compute is within the combine?
The value a plan carries and the margin it retains are not often the identical quantity as soon as inference and serving prices are counted. Margin is the place most groups level Genie One first, and it’s why design software program, market, and shopper web firms are standing it up because the self-service entry level for his or her finance groups.
› The place is consumption income in danger throughout blended subscription and utilization pricing?
Consumption you possibly can meter accurately is income you possibly can acknowledge, and the worth is in catching a spot because it varieties, not on the shut. A digital wagering firm has taken this the furthest: it changed fragile spreadsheet forecasting with ruled income forecasting fashions, added liquidity evaluation, and gave its management group a Genie Considered one of its personal.
› The place is compute spend susceptible to outpacing the runway as utilization grows?
The time to behave on burn is earlier than the subsequent compute dedication is signed. A digital asset platform runs its monetary money pipelines and blockchain analytics on the platform for that reason, and a world supply platform runs finance forecasting, finance apps, and workflow automation the identical manner.
And since each determine traces to its supply, each permission holds, and the price of the AI itself stays ruled underneath one mannequin, the reply is one finance can act on. Genie One readies the transfer, whether or not that could be a repricing, a metering repair, or a slower drawdown, and an individual within the loop makes the decision.
From first reply to finance platform
The arc throughout tech and AI-native firms is constant. Finance lands with ruled, self-service solutions. Then it begins consolidating. An AI-native software program chief that runs finance evaluation stay on Databricks is now validating Genie One for finance and operations. A monetary software program chief is consolidating its finance reporting stack and constructing brokers for its finance group. Others are retiring the purpose instruments finance accreted over a decade: dashboards on legacy BI platforms, desktop data-prep workflows numbering within the lots of, finance apps sitting on a separate cloud knowledge warehouse. Every migration places extra of the enterprise’s which means into the ontology, and Genie One’s solutions get higher with it. Finance finally ends up operating on the identical platform because the product: one platform for knowledge and AI that scales from pre- to post-IPO with out re-platforming.
A knowledge-smart AI coworker constructed for the best way finance works
That is what finance departments have been asking for: an AI coworker that stays present, stays ruled, and retains studying the enterprise. It is usually how we run our personal. Michael Schaaf, Senior Director of Finance at Databricks, confirmed how the Databricks finance group runs on Databricks within the Databricks on Databricks webinar, and brings the total story to Knowledge + AI Summit within the session How Databricks Makes use of Databricks to Run Its Personal Finance Group, and Why That Adjustments All the things.
Tech and AI-native firms will continue to grow on AI. Genie One is how finance protects the unit economics that development is determined by. See what a data-smart AI coworker appears like for a finance group: databricks.com/product/ai-bi/genie
Regularly requested questions
What’s altering for finance in an AI-native enterprise?
Extra of the selections that transfer unit economics (compute, pricing, and income recognition) are made by brokers. Finance’s mission to guard the margin has not modified. The velocity and complexity of what it has to know and govern has.
Does Genie One make pricing, packaging, or compute choices?
No. These calls belong to product, engineering, and income accounting. Genie One offers finance an correct, ruled view so a forming danger will get seen early and routed to the proprietor who acts on it.
Why do ontology and governance matter to a tech firm CFO?
Ontology captures what the numbers imply for your online business and retains it present, so a solution is appropriate and never simply correct. Governance retains each determine traced, permissioned, and cost-controlled. Collectively, they make a solution protected to behave on.
How is Genie One completely different from an AI dashboard or BI software?
A dashboard reveals you what the info says. Genie One helps you act on it: grounded in your ontology, ruled finish to finish, with an individual making the ultimate name.
