Andera secures $37 million to enhance audit automation through AI technology.

      Auditing is an aspect of corporate life that rarely crosses people's minds, yet nearly everyone ends up bearing its cost. Public companies devote thousands of hours annually to gather evidence, assess controls, and document their findings. Much of this process is still performed manually, using Excel, much like it was in 2002.

      Andera aims to shift this labor to AI. The San Francisco-based startup, established in 2024, has secured a $37 million Series A investment led by Lightspeed Venture Partners to fully automate the internal audit process.

      A Problem of Reasoning, Not Rules

      The resistance to automation in auditing is understandable. Auditors must interpret complex, multi-format evidence, assess whether it meets control requirements, and document everything sufficiently to withstand regulatory scrutiny—a task no rules engine has successfully accomplished.

      Large language models may provide a solution. Andera’s platform can analyze financial evidence from spreadsheets, PDFs, screenshots, and journal entries, delivering an audit determination for controls per Sarbanes-Oxley and other regulations. “The models are already quite intelligent,” said CEO Aryo Patel. “The challenging part is narrowing down from billions of tokens to the 30,000 that are relevant.”

      Patel, who has previous experience at Microsoft and Jane Street, and co-founder Tinah Hong, formerly with Stripe, met in middle school in Chicago. They have combined engineers with experienced auditors, including a former Deloitte accountant, to develop their product.

      The Dilemma for the Big Four

      This pitch comes at a timely juncture. According to Lightspeed, the average public company allocates around $3 million per year for audit fees, with Fortune 100 companies spending over $20 million. Financial leaders are now demanding more efficiency from their back office operations.

      This situation places the Big Four accounting firms in a difficult position. Their business model relies on billing by the hour, so the more effective the AI they create, the more it threatens their own income. This dynamic mirrors the pressure currently affecting the consulting industry and creates an opportunity for a startup that has nothing to lose.

      Targeting White-Collar Professions

      Though Andera is small, with only a few employees, it claims to already serve Fortune 100 clients. It is part of a rising trend of startups targeting AI not toward chatbots but toward high-earning professions, including corporate lawyers, patent attorneys, and now auditors.

      The reassuring narrative, as presented by Lightspeed, suggests that auditors will be augmented rather than replaced, allowing them to be liberated from tedious busy-season tasks to focus on real risks. Conversely, the less comforting thought now weighing on every white-collar worker is: what occurs when AI progresses to the point of executing the judgment tasks as well?