startups

Rillet’s $100M Round Is a Warning Shot for Oracle and NetSuite

A finance team of three helping run $2B in revenue is either the future of business—or a bloody good reason to audit every AI claim twice.

Rillet’s $100M Round Is a Warning Shot for Oracle and NetSuite

A finance team of three helping run $2 billion in revenue is either the future of business—or a bloody good reason to audit every AI claim twice.

Either way, Rillet just raised $100 million at a $1 billion valuation because investors think the old way of running finance is finished. Not improved. Finished. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

Rillet did not raise money. It sold a verdict.

On August 19, AI-native accounting platform Rillet announced a $100 million Series C led by ICONIQ, valuing the company at $1 billion. Sequoia, Andreessen Horowitz and a long list of existing backers joined in. It was Rillet’s third funding round in a year and took its total capital raised past $200 million. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

That alone is a decent startup story. The more interesting bit is how fast the round came together. CEO and co-founder Nicolas Kopp told TechCrunch that Rillet had not gone out to raise; after a board meeting, investor conversations turned into a unicorn round in less than 48 hours. That sort of speed is usually reserved for a hot consumer app, a defence company with government contracts, or a proper AI frenzy. Not bookkeeping. ([techcrunch.com](https://techcrunch.com/2026/08/21/how-ai-accounting-startup-rillet-raised-100m-and-became-a-unicorn-in-48-hours/))

But calling this “bookkeeping software” misses the point.

Rillet is trying to replace the general ledger—the central system a company uses to record transactions and produce financial statements—with software built around real-time data, AI agents, human approvals and an audit trail. Its pitch is blunt: legacy enterprise resource planning systems, or ERPs, are databases built for a batch-processing era. They record what happened. Rillet wants the ledger to become the place where humans and software agents actually do the work. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

That is why Oracle, NetSuite, SAP, Workday, Sage and Microsoft ought to pay attention. Rillet says customers are replacing those systems, not merely buying an AI add-on beside them. It reports more than 600 customers, doubled new annual recurring revenue in the prior three months, and says it is expanding beyond tech into biotech, healthcare, fintech, logistics and professional services. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

The company’s claims deserve the usual dose of investor scepticism. Every founder makes the future sound inevitable after taking a cheque. But the direction of travel is hard to argue with: companies are becoming less patient with paying armies of smart people to copy figures between systems, reconcile accounts and wait weeks to see what is happening in their own business.

The real product is not AI. It is permission to trust it.

Most AI accounting pitches sound like a flashy spreadsheet with a chatbot bolted on. That is not enough. Finance is where optimism goes to get mugged by reality.

If an AI tool writes a duff marketing email, you look silly. If it misclassifies revenue, stuffs up a reconciliation or produces a number that cannot survive an audit, you have a serious problem. Potentially a board problem. Potentially a regulator problem. Anyone selling “fully autonomous finance” without talking about controls is selling a fantasy with a nice landing page.

Rillet’s more credible angle is that the agents are meant to operate within the ledger itself, on live structured data, with shared policies, human approval authority and a complete audit trail. The company says that design gives finance teams visibility into what the software did and why. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

This is also why its April alliance with EY matters more than the usual corporate-partnership press release. EY said the relationship combines Rillet’s AI-native platform with finance transformation, risk and controls expertise. The stated work includes journal entries, reconciliations, revenue recognition, close management and reporting. In plain English: the startup is trying to get credibility in the bits of finance where a cock-up costs real money. ([ey.com](https://www.ey.com/en_gl/newsroom/2026/04/ey-announces-alliance-with-rillet-to-provide-ai-native-finance-transformation-with-risk-and-controls-built-in))

That is the moat people overlook. The model is not the moat. Everyone has access to models, or soon will. The hard part is embedding intelligence into ugly, high-consequence workflows without wrecking governance. You do not beat Oracle by generating prettier dashboards. You beat it when a chief financial officer believes your system can close the books, survive an audit and tell the truth under pressure.

Why the incumbents are vulnerable—and why they are not dead

A lot of people get carried away at this point and declare a “SaaSpocalypse.” Relax.

Oracle, SAP, NetSuite and the rest have massive installed bases, deep integrations, enterprise sales machines and customers who regard changing a general ledger as open-heart surgery. Rillet’s early investor, Sequoia’s Julien Bek, made exactly that point in 2025: replacing the ledger is difficult because it is so central to a company’s financial life. ([techcrunch.com](https://techcrunch.com/2025/05/28/rillet-raises-25m-from-sequoia-to-automate-general-ledger-systems-using-ai/))

The migration pain is real. Data is messy. Policies are inconsistent. Years of exceptions have been buried in spreadsheets and carried around in the heads of two long-serving finance people who are already overworked. A shiny AI interface does not solve that by itself.

But incumbents have a structural problem: their systems were designed around periodic updates, cumbersome implementations and human operators shuffling work across spreadsheets and bolt-on tools. Rillet argues a new architecture can shorten implementation to weeks rather than months, while keeping the ledger continuously current. That is the sort of claim buyers must verify in their own environment—not take on faith—but it identifies the right battlefield: time to value, not feature-count bingo. ([rillet.com](https://www.rillet.com/blog/rillet-raises-70m-series-b-from-andreessen-horowitz-and-iconiq))

Here is the second-order implication. If a finance team can operate from a continuously updated ledger, “month-end close” starts looking like an absurd ritual. Management does not need a rear-view mirror delivered 15 business days late. It needs a live dashboard of cash, revenue, margin, tax exposure and collections while there is still time to act.

That changes more than the finance department. It changes how a founder runs the company. Pricing errors get spotted faster. Bad customer cohorts get noticed earlier. Spending decisions get less theatrical because the numbers are not ancient by the time they reach the room.

The overlooked angle: this is a labour shortage trade, not merely an AI trade

The lazy story is that Rillet wants to sack accountants. The better story is that accounting has a supply problem and too much good talent is trapped doing low-value work.

The US Bureau of Labor Statistics projects employment of accountants and auditors to grow 5% from 2024 to 2034, adding about 72,800 jobs. Its view is not that automation eliminates demand; rather, routine work can shift accountants toward analysis and advisory duties. ([bls.gov](https://www.bls.gov/ooh/Business-and-Financial/Accountants-and-auditors.htm))

That makes sense to me. Good accountants are not data-entry clerks with a CPA. A proper finance person helps you decide whether growth is profitable, whether cash is real, whether a supplier deal is dodgy, and whether you are about to create a tax headache that will ruin your weekend.

Rillet’s investor materials claim that Mercor has used its agents to scale beyond $2 billion in annual recurring revenue with a three-person finance team. Treat that as a company claim, not scripture. But even if the exact example is exceptional, the principle is dead right: the prize is not fewer capable people. The prize is fewer capable people wasting their brains on reconciliation theatre. ([businesswire.com](https://www.businesswire.com/news/home/20260819978953/en/Rillet-Raises-%24100M-Series-C-at-%241B-Valuation-to-Build-Accounting-Superintelligence))

The contrarian point is this: AI will not make finance less important. It will make finance more important because the businesses that trust their numbers in real time will move faster than those waiting for a monthly post-mortem.

What this means for you

If you are a founder or operator, do not rip out your accounting system on Monday because a startup raised $100 million. That is how people create six-figure cleanup bills.

Do this instead:

1. Measure your current finance drag. How many days does close take? How many reports are manually rebuilt? How many key numbers live in a spreadsheet because nobody trusts the ERP? Write down the actual hours and cost.

2. Pick one painful workflow, not “AI transformation.” Start with bank reconciliations, revenue recognition checks, board reporting, collections or spend categorisation. A narrow workflow with a measurable before-and-after beats a grand strategy deck every time.

3. Demand an audit trail before you admire a demo. Ask: What did the agent do? What data did it use? Who approved it? Can we reverse it? If the vendor cannot answer in plain English, do not let it near your ledger.

4. Keep humans responsible for judgement. Automate the grind. Keep sign-off, exceptions, policy decisions and material risk with people who understand the business. Faster nonsense is still nonsense.

5. Use your finance stack as a competitive weapon. The aim is not to have fashionable software. The aim is to know your cash, unit economics and problems earlier than your competitors do.

Rillet’s $100 million round is a bet that the ledger becomes an operating system rather than a digital filing cabinet. The valuation may prove too rich. Some promises will certainly be overcooked. But the underlying shift is real: companies that still treat finance as a monthly administrative chore are leaving speed, control and money on the table.

That is a stupid place to leave it.

Sources