
Like most of his fellow accountants, Jin Chang tried to quit the industry several times because of the exhaustion of the grunt work. That is until 2020, when he had the idea to build software that could automate more parts of the job at a time when the accounting profession is hemorrhaging talent. The result, Fieldguide, predates the rise of ChatGPT, although its artificial intelligence approach has been fueled by the success of generative AI.
On Monday, Fieldguide announced a $75 million funding round led by growth equity group Goldman Sachs Alternatives, valuing the company at $700 million, with participation from new investor Geodesic and existing investors Bessemer Venture Partners, 8VC, and Thomson Reuters.
“Audits, on the surface, can be boring, but actually the purpose behind what they do is noble and important for the business world,” Chang said. “We’re replacing a fragmented technology stack.”
The future of accounting
As Chang notes, the certified public accounting, or CPA, industry is going through a talent crisis. “And it just got worse,” he said luck. “I would call it existential.” The number of candidates writing the CPA exam is at a 17-year low, with more than 75% of today’s CPAs expected to retire in the next decade, according to Fieldguide. The work of accounting firms, however, is important for many large businesses, such as auditing quarterly financial reports made by public companies.
Chang’s previous job was at the AI-powered legal firm Atrium, created by Twitch cofounder Justin Kan, which closed in 2020. Chang had a similar idea to use AI to automate accounting work, but decided to take a software approach rather than building a next-generation accounting firm. Many of the big companies, he argues, have decades-long relationships with their accounting firms. Instead of rebuilding that trust from square one, he can now work with the accounting firms himself.
Fieldguide initially struggled to find traction with its AI-native approach. In the long run, however, its decision to build the business for a world of native AI has been proven by the sudden popularity of several language models from OpenAI and Anthropic. Over the past 18 months, Chang says Fieldguide has begun to achieve product market relevance.
“That’s when we started to feel the CPA industry as a whole was not only receptive, but there was a strong appetite,” he said. Today, Fieldguide is used by half of the top 100 US accounting firms, including KPMGdespite varying degrees of adoption of his agentic tools.
According to Chang, Fieldguide’s AI agents are replacing functions that are usually outsourced to offshore companies in countries such as India and the Philippines, effectively acting as extended teams of CPAs in the US That may include testing revenue figures in financial statements, planning how to conduct an audit based on a client’s industry, and writing procedures that an audit team must go through. “Instead of offshoring the work, our AI agents execute the first job pass, then people come in to check the results,” Chang said. “What we leave for the expert people is to interpret the results, and also, give up time so they can do more work with added value, which builds trust in client relationships.”
Chang said that with Fieldguide, a traditional 10-person audit team may be able to audit three people within the next few years, and that soon, its agents will be able to conduct a higher level of strategic thinking. Rather than replacing jobs, he argued that automation could help junior auditors accelerate their careers. “We will see less burnout in the industry, because the work will be interesting at a higher level.”
Goldman’s AI bet
Darren Cohen, the global co-head of growth equity at Goldman Sachs Alternatives, said his team is tracking a variety of verticals that could be an interesting fit for AI, including accounting. They tracked Fieldguide for about two years before deciding to become the lead investor for its Series C, in part after spending time talking to Fieldguide customers. “Clients are getting a 30-40% improvement in efficiency,” he says luck. “It’s a great case study in AI applications. These aren’t just big language models – they’re purpose-built models designed for specific use cases.”
Cohen argues that Fieldguide’s pre-ChatGPT history, at a time when many new startups were racing to use the novel technology, helps explain its success. “They’re deep in the domain, and that’s why they get that advantage” he said, adding that Chang’s background in accounting gives him a “primary understanding” of the industry’s inefficiencies. “You want people who are deeply committed to the mission they’re pursuing.”
Fieldguide currently has 160 employees with plans to double in the next year. The company is now 25% former auditors, according to Chang, and continues to increase the number of engineers and CPAs. “We combine the best of Silicon Valley AI engineering and subject matter expertise directly from the field,” Chang said.
Although a strategic acquisition of an accounting firm may seem like the most logical outcome for Fieldguide, Cohen said he hopes it will achieve an initial public offering. “My ideal world is for them to achieve their vision,” he said. “They expanded it into all these different disciplines: auditing, tax, accounting. Eventually it became a very big business, and they could take it public when it was ready to go public.”






