Rillet raises $100m in 48 hours after a board meeting

Rillet closed a $100 million round at a $1 billion valuation two days after a routine board meeting, chief executive Nicolas Kopp said in an interview with TechCrunch.

The company was not raising. The numbers in the room did the asking.

At that meeting, Rillet took investors through its progress since a $70 million Series B last summer. Kopp said annualised revenue rate had doubled in the previous quarter alone.

He also reported new clients, several of them public companies, and an alliance with EY to put AI tools into the auditing firm's work.

Texts and calls followed. Forty-eight hours later the round was done, led by Iconiq, with Sequoia among the investors returning.

Iconiq general partner Seth Pierrepont, who co-led the Series B and now joins the board, told TechCrunch the speed was misleading. Sequoia's Julien Bek, who led the Series A last summer, said the firm already had the context it needed.

That is the mechanic worth noting. Neither lead was doing diligence in 48 hours; both had been watching the company for a year and were re-upping on a number they had been tracking.

Rillet came out of stealth two years ago and has now raised $200 million in total, from backers including Iconiq, Andreessen Horowitz and Sequoia.

It has 600 customers, Kopp said, most of them replacing incumbent systems rather than trialling a second one alongside them. He put the split at 50% coming from Intuit, 30% from NetSuite and Sage Intacct, and 20% from Oracle, SAP, Workday and Microsoft products.

The customer list runs from laundromats to a major sports franchise, on his account.

The product is built for AI agents working alongside human accountants. Customers can route requests to the foundation model of their choice, and Kopp said Rillet's setup stops those models training on customer data, with no cross-training between accounts.

About three months ago the company shipped a governance feature that lets accountants audit every decision an agent has made, including which numbers it pulled and how it calculated them. Kopp said compressing agent data into something a human could read was the hard part.

Regulation still assumes a person is in the loop: rules for public companies require a human to approve every transaction an AI agent makes.

The demand behind all this is a labour shortage. The Controllers Council Organization found in a recent report that 61% of finance leaders had struggled to hire finance, accounting and CPA talent in the past year, and the number of US accounting graduates has been falling since at least 2010.

The Bureau of Labor Statistics projects accounting-related roles growing at least 5%, adding 72,800 jobs by 2034, and does not expect AI to cut demand.

For founders, the lesson is less about the 48 hours than about what preceded them. The round closed fast because two existing investors had watched the metric move for four quarters and needed no persuading when it doubled. Inbound leverage is built in the year before the board meeting, not in the meeting.

More: Rillet