Every year, audit teams at public companies spend thousands of hours collecting evidence, testing controls, and producing workpapers, most of it manually, most of it in Excel, and most of it in exactly the same way it was done when Sarbanes-Oxley was enacted in 2002. For the average public company, audit fees cost $3M annually. For Fortune 100 companies, they spend north of $20M annually. For internal audit, SOX controls are just a small portion of an enterprises’ controls universe. As businesses became more complex over the years, what was historically a SOX-and-compliance function is now the last internal line of defense critical to investor protection and enterprise integrity. Despite the function’s critical role, the experience of internal audit hasn’t been reimagined in two decades.
That changes today. We’re thrilled to announce that Lightspeed is leading a $37M Series A in Andera, the company building the AI-native platform to automate internal audit.
Why Audit? Why Now?
It was clear to our team that audit is a consequential category for AI and spent several months building a thesis here before writing a check. We spoke with finance and audit leaders across Fortune 500 companies spanning technology, financial services, healthcare, and consumer. We evaluated early-stage players building in the space. We talked to Big Four partners, former PCAOB officials, and practitioners across the industry. What we found was a market at a pivotal moment.
The core workflow of audit has resisted automation for decades because it is fundamentally a reasoning problem, not a rules problem. An auditor must read complex, multi-format evidence, reason about whether it satisfies a control objective, and produce documentation that holds up to regulatory scrutiny. No rules engine could handle that. But the convergence of large language model reasoning, long context windows, and agentic tool use has changed the calculus. As Aryo Patel, Andera’s CEO, put it: “The models are already sufficiently smart. The hard problem is getting from billions of tokens to the 30,000 that matter.”
At the same time, cost pressure has become a mandate. CFOs at publicly traded companies are demanding 200–300% efficiency gains from back-office functions. The Big Four, whose business model is built on billing hours, face a genuine innovator’s dilemma: they cannot build the best AI product without cannibalizing their own headcount. They will partner with the companies building at the cutting edge of innovation. That window rarely exists for a new entrant, and it is open right now.
We evaluated every company going after this problem. On technical architecture, breadth of vision, and institutional credibility, Andera was the clear outlier.
What Makes Andera Different
Today, Andera automates every control (SOX, operational, compliance, etc.) test end-to-end from user access provisioning to even the most complex bad debt reserve and management review controls. The platform processes hundreds of millions of tokens of financial evidence per control across Excel workbooks, PDFs, system screenshots, journal entries and applies AI to identify the right documentation and render an audit judgment. However, the future of Andera expands to documenting, assessing and testing any type of control irrespective of the industry, complexity or system. Andera’s key differentiator is how they have built the product from the start to reflect the breadth and depth of an auditor’s work.
Andera’s architecture is built to reflect a seasoned auditor’s intuition. The team has built a best-in-class custom engine that handles how an agent reads, processes, and writes back into Excel dynamically the same way auditors navigate their workpapers. What is under the hood is a proprietary agentic retrieval that pulls the right evidence across billions of tokens without vector-retrieval’s recall failures, probabilistic beam search that up-samples compute in uncertain regions of the data graph, and custom agents that manipulate financial workpapers with a precision that is required to instill trust among the Fortune100 customers they work with today.
The Team That Earned Our Conviction
We have known the co-founders, Aryo and Tinah, since the formation of Andera. The conviction we have in this team isn’t the result of a few diligence calls. It has compounded over time, watching them build, make hard decisions, and consistently execute.
Aryo Patel (CEO and Co-founder) came to this problem from the intersection of finance and AI infrastructure, with stints at Microsoft on Azure infrastructure scaling and Jane Street on the trading desk. He thinks about audit not as an IT challenge but as a reasoning problem that happens to live inside finance. That reframe is the foundation of everything Andera is built upon.
Tinah Hong (CTO and Co-founder) has been close friends with Aryo since middle school in Chicago through their MIT undergrad days, where the core of this founding partnership was built long before the company was. Tinah formerly was at Stripe building machine learning systems and Microsoft as a software engineer. Andera’s architectural decisions that make Andera’s testing accuracy possible are a direct product of her technical judgment.
The team has paired their extensive technical depth with the best industry expertise. They brought on seasoned experts that lived the problem in: Carina Averilla, their first hire, was an ex-Deloitte auditor and former Head of US Accounting at Revolut; Jared Lauber, who leads Partnerships, has spent 30 years as a risk and audit leader at places like EY, Instacart, McKesson, and Williams-Sonoma.
What Great Auditors Were Meant to Do
The best audit executives didn’t get into this field to spend their careers sampling populations in Excel. They got into it to help their organizations see around corners, to identify where the real risks are, where growth is leaving money on the table, and how to build resilient enterprises. The tension of the current model is that the most intellectually demanding parts of the job get crowded out by the operational grind of getting through busy season.
Andera changes that equation. When evidence collection, control testing, and workpaper generation are handled by AI that is more consistent, more thorough, and faster than any human team, audit professionals get something back that the industry has been slowly taking away: time, attention, and the ability to think. The auditor of the future isn’t displaced by this technology. They are extended by it, able to see more of the enterprise, reason about more of the risk surface, and contribute to decisions that actually move the business forward.
That is the vision Andera is building toward. Andera is more than a tool that assists with audit. It is the AI that thinks alongside the best audit minds in the world, calibrated to each enterprise’s specific risk appetite, control environment, and strategic context, so that the humans in the room can do what they were actually hired to do.
We believe that is one of the most important transformations happening in enterprise software right now. We’re proud to lead this investment and build it alongside Aryo, Tinah, and the Andera team.
The content here does not constitute an offer to sell or a solicitation of an offer to buy any securities or investment advisory services.
The views expressed are those of the authors and do not necessarily represent the views or opinions of Lightspeed. Other market participants could take different views. Unless otherwise indicated, the inclusion of any third-party firm and/or company names, brands and/or logos are for representational purposes and does not imply any affiliation with these firms or companies and also does not imply their endorsement of the views expressed by the authors.
Certain information contained herein is based on information from various sources prepared by third parties. While such sources are believed by Lightspeed to be reliable, neither Lightspeed nor its affiliates assume any responsibility for the accuracy or completeness of such information, and such information has not been independently verified by Lightspeed. For more details please see https://www.lsvp.com/legal.
Authors
