Andera secures $37 million to automate auditing using AI.

      Audit is an aspect of corporate life that often goes unnoticed, yet it incurs significant costs for almost everyone. Public companies dedicate thousands of hours annually to gather evidence, test controls, and document their findings, primarily using Excel, much like they did two decades ago in 2002.

      Andera aims to transfer this labor to AI. Founded in San Francisco in 2024, the startup has secured a $37 million Series A funding round led by Lightspeed Venture Partners to completely automate the internal audit process.

      A reasoning challenge, not a rules issue

      Audit has been resistant to automation for a reason. An auditor must analyze complex, multi-format evidence, determine if it meets a control standard, and document everything sufficiently to withstand regulatory scrutiny, which no rules-based system can adequately accomplish.

      However, large language models have the capability to try. Andera's platform can interpret financial evidence from spreadsheets, PDFs, screenshots, and journal entries, then deliver audit assessments for controls related to Sarbanes-Oxley and other regulations. "The models are already quite advanced," stated CEO Aryo Patel. "The challenging task is narrowing down from billions of tokens to the 30,000 that are crucial."

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

      The dilemma for the Big Four

      This proposition arrives at a timely juncture. According to Lightspeed, the average public company allocates around $3 million annually for audit fees, while a Fortune 100 company exceeds $20 million. Finance leaders are now pushing for their back offices to accomplish much more with fewer resources.

      This scenario creates a dilemma for the Big Four accounting firms. Their revenue model is based on billable hours, so as they create more effective AI, it could reduce their own income. This is the same pressure currently affecting the consulting sector, creating an opportunity for a startup without existing revenue to disrupt.

      Targeting white-collar sectors

      Though Andera is still small, with just a few employees, it claims to already be collaborating with Fortune 100 clients. It is part of a growing wave of startups that are focusing AI efforts on lucrative professions, such as corporate lawyers, patent attorneys, and now auditors.

      The optimistic view, which Lightspeed promotes, is that auditors will be augmented rather than replaced, given more freedom from repetitive tasks during peak seasons to concentrate on genuine risks. The more concerning perspective, however, is the question that every white-collar worker is starting to consider: what will occur when AI becomes advanced enough to handle the decision-making as well?