Basis Builds a Domain-Specific AI Assistant for Accounting Firms

01 /The Concept

Accounting firms are facing a staffing crisis that looks a lot like a technology opportunity. Basis is building the AI platform designed to fill the gap – and has raised $3.6M at pre-launch stage to pursue it.

The platform is purpose-built for accounting firms that manage books for other companies. Its AI operates at several levels simultaneously. The conversational layer handles finance and accounting queries using domain-specific training and authoritative reference data rather than general-purpose generation – which matters because hallucination in models like ChatGPT makes them unsuitable for tasks where factual accuracy is non-negotiable.

More interesting is the transaction-level work. Every incoming transaction in a client's books needs to be correctly categorized. Basis can explain what a transaction likely represents – identifying, for instance, that an $8,000 transfer to an advertising firm in August almost certainly funded a marketing campaign, which means it should be coded to the appropriate expense category and the firm should request a work confirmation from the vendor. It can also pull from client contracts with suppliers and customers to bring forward context about scheduled payments and required documentation.

At period close, Basis reviews the entire ledger before it's signed and filed – flagging inconsistencies, gaps, and contradictions, and producing a suggested action list to resolve them. When clients send questions about their own financials – the inevitable "why did this happen" and "what does this mean" emails – Basis drafts responses using both general accounting knowledge and the specific data of that client's account.

The platform is in pilot with early clients, not yet publicly available.

02 /Why It Matters

Benjamin Franklin's observation that only death and taxes are certain has acquired a new commercial dimension: the infrastructure around tax compliance is large, distributed, and now structurally understaffed. There are 88,500 accounting service businesses in the US alone.

The workforce data is stark. More than 300,000 accountants and auditors left their positions in the US in 2020–2021. Among people who had been in the profession for more than six years, departure rates jumped from 63% in 2015 to 82% by late 2023. The Wall Street Journal recently ran a piece on the difficulty of attracting new entrants to the field. What's replacing experienced professionals is a cohort of newer staff who, without support, can't sustain the same output.

The consequences show up in client-facing metrics. Firms with under $8M in revenue were forced to turn away 6.46% of clients in 2022. Even the largest firms – with over $300M in annual revenue – declined 2.4% of potential engagements due to capacity constraints. Demand is high; labor supply is not. That's a structural condition, not a temporary one, and it points toward AI augmentation as the primary lever available to the industry.

The broader pattern Basis represents is the proliferation of vertical AI platforms using Retrieval Augmented Generation – anchoring AI output to curated, verified data sources rather than relying on general training. Similar platforms have emerged in logistics (Rippey.AI, $4.8M), startup finance advisory (Mantle, $7.68M), and travel planning (Mindtrip, $7M). Accounting is a logical candidate: it's terminologically dense, high-stakes on accuracy, and has enormous quantities of structured historical data to ground AI responses.

03 /Opportunities

The general direction – vertical AI assistants for specific professional niches – is clear. The case for urgency is also clear: the window for establishing a position in any given niche is open now and closing as funding flows in.

The non-obvious selection criterion is what makes this review worth reading carefully. The most commercially promising niches for vertical AI aren't the most technically exciting or the most glamorous. They're the ones where attrition is rising and new entrant flow is falling – where AI can turn a deteriorating situation into a manageable one rather than merely making a good situation marginally better. Accounting fits that description precisely.

The question is which other professional categories share the same profile: structurally important work, growing demand, a workforce that's leaving faster than it's being replaced, and no incumbent AI platform that already owns the space. Those are the niches worth entering now. They may not generate compelling pitch deck slides, but the demand is real and the competition is thinner precisely because the work is unglamorous.