AI reshapes accounting firms but full automation lags

A blog post by accounting technology commentator Jason Staats has set out a shortlist of artificial intelligence tools that he said accounting firms now use for core work, tax research, client intake, meeting notes and internal automation.  Staats is a CPA, a firm owner, a YouTuber, a podcaster, and the founder of an accounting community called Realise.

Staats described a split between products built specifically for accounting firms and general-purpose AI products that firms have adopted for finance operations, document review and staff productivity.

He also argued that accounting has not seen full automation because much of the remaining work sits in the gaps between systems, vendors and processes, rather than inside a single ledger or tax engine.

Ledger alternatives

In his overview of accounting-specific products, Staats pointed to Digits and Kick as "full ledger replacements" for firms with cash basis clients. He positioned them as alternatives to QuickBooks and Xero.

He also highlighted Blue J for tax research and drew a contrast with tools sold by established tax software suppliers.

Staats said firms also increasingly use automation agencies. He described them as service providers that build workflows and automations for accounting practices.

Firm essentials

Staats said every firm should now use AI meeting notetakers. He described tools that create transcripts of client calls and internal discussions. He also said the market now includes products designed for accounting firms and products that integrate with practice management systems.

He named Vinyl, Ping and Abacor as examples in that category.

On personal tax workflow, Staats singled out what he called "a 1040 intake solution". He described it as a major shift in tax technology.

He cited StanfordTax, Soraban and Truss as current leaders. He said his view reflected feedback from accounting firms in a private group.

Staats also recommended a paid version of OpenAI's ChatGPT that he said includes protections on data handling that do not apply to free or consumer plans.

"The only version that comes with the data protections you need (besides the more costly Enterprise plan) is 'ChatGPT Business'," said Staats.

Wider AI stack

Beyond accounting-specific tools, Staats listed a set of general AI products that he said firms have started to use across finance and operations.

He highlighted Ramp's AI Policy Agent. He described a workflow where a company provides its expense reimbursement policy and the system approves, denies, or flags submissions for review. He also wrote that the tool points users to relevant policy lines.

For accounts payable, Staats named MakersHub. He described it as an AI accounts payable platform. He compared it with bill pay features inside banking platforms and with Ramp's bill pay product.

He also referenced Google's NotebookLM and a feature he called "Audio Overview". Staats wrote that users can upload monthly financials and general ledger information and receive an audio summary presented as a conversation.

Staats also recommended speech-to-text products. He named Wispr Flow for PC and Superwhisper for Mac.

He also listed Lovable. He described it as a way to build apps without writing code.

Website positioning

Previously, Staats also set out what he described as three common mistakes on accounting firm websites. The first focused on design, which he said often takes priority over clarity and conversion. The second covered language, which he said too often reflects internal jargon rather than client phrasing.

He suggested firms use meeting transcripts to identify the words clients use when they describe services. The third mistake covered the next steps. Staats said firm websites should make it clear what the visitor should do next.

Why automation stalls

Staats argued that accounting still contains work that is hard to standardise. He said many repetitive tasks have already been automated. He described what remains as messy work across systems, data sources and human behaviour.

He gave examples such as chasing information from business owners, transferring data between systems, and downloading transactions from banks that do not support data feeds.

Staats said firms should look at what parts of their work resemble transaction classification, bill entry, or payroll processing. He argued that those activities could diminish significantly over time.

He also argued that bookkeeping has shifted away from data entry towards oversight and control. He said clients still bring firms unfamiliar forms and expect help navigating compliance.

Staats also linked technology progress with rising complexity elsewhere. He said tax computation software coincided with more complex rules, and eCommerce convenience increased purchasing volumes.

"Think buying AI apps for your firm is expensive? Consider the cost of having an AI-inept workforce in 3 years,"  Staats wrote.

He said the industry should expect continued investment and experimentation in tools such as tax intake, as vendors approach the same workflow in different ways.