Andera raises $37M for AI-powered audit automation
▼ Summary
– Andera, a 2024 San Francisco startup, has raised a $37mn Series A to automate internal audit using AI, processing evidence like spreadsheets and PDFs to render audit judgments.
– Audit has resisted automation because it requires reading messy, multi-format evidence and making judgments, which rules engines cannot do, but large language models can now attempt this.
– The Big Four accounting firms face a dilemma: automating audit with AI would cannibalize their revenue from billing hours, creating an opening for startups like Andera.
– Andera works with Fortune 100 customers and is part of a wave of startups targeting well-paid professions like auditors, corporate lawyers, and patent attorneys with AI.
– The optimistic view is that AI extends auditors by freeing them from grunt work, while the concerning question is what happens when AI becomes capable of making judgments itself.
Most people never think about corporate audit until they see the bill. Public companies pour thousands of hours each year into gathering evidence, testing internal controls, and documenting everything for regulators. The vast majority of that work still relies on manual effort in Excel, using methods that have barely changed since 2002.
Andera is betting that those days are numbered. The San Francisco-based startup, founded in 2024, has secured a $37 million Series A led by Lightspeed Venture Partners to build an AI-powered platform that automates the entire internal audit process.
A reasoning problem, not a rules problem
Audit has long resisted automation for a good reason. An auditor must parse messy, multi-format evidence, determine whether it satisfies a specific control, and document the reasoning clearly enough to withstand regulatory scrutiny. No traditional rules engine could handle that.
Large language models offer a new path. Andera’s platform ingests financial evidence from spreadsheets, PDFs, screenshots, and journal entries, and then produces an audit judgment for controls under Sarbanes-Oxley and other frameworks. “The models are already sufficiently smart,” said CEO Aryo Patel. “The hard problem is getting from billions of tokens to the 30,000 that matter.”
Patel, a former Microsoft and Jane Street engineer, co-founded the company with Tinah Hong, a Stripe alum. The pair met in middle school in Chicago. They have combined engineering talent with career auditors, including a former Deloitte accountant, to shape the product.
The Big Four’s dilemma
The timing works in Andera’s favor. Lightspeed notes that the average public company spends roughly $3 million annually on audit fees, while Fortune 100 firms pay more than $20 million. CFOs are under growing pressure to do more with leaner back offices.
That creates a real problem for the Big Four accounting firms. Their business model depends on billable hours, so the better their internal AI tools become, the more they risk eating their own revenue. It is the same structural squeeze now reshaping the consulting industry, and it leaves a clear opening for a startup with no legacy revenue to protect.
Coming for the white-collar professions
Andera is still small, with only a handful of employees, but it claims to already work with Fortune 100 customers. It joins a growing wave of startups targeting not chatbots or customer service, but the well-compensated professions: corporate lawyers, patent attorneys, and now auditors.
The optimistic story, the one Lightspeed is telling, is that auditors will be extended, not replaced , freed from the drudgery of busy-season grunt work to focus on genuine risk. The less reassuring version is the question every white-collar professional is now learning to ask: what happens when the AI gets good enough to handle the judgment too?
(Source: The Next Web)
