One firm gave AI tools to its whole workforce and ended up with more than 2,500 agentic workflows, including 500 tools just for reviewing board meetings. Caseware’s Jason Bradley told that story to AccountingWEB last week to show how far audit has moved on: firms are now picking two or three specific bottlenecks and demanding traceability from every tool. That lesson reaches well beyond audit. The same test, whether an answer can be proved or merely sounds right, is now being applied to accounting firms themselves every time a prospect asks an AI assistant who to hire.

The bigger picture: the profession has stopped being impressed by plausible

The first wave of AI in accountancy was about access. Give people the tools and see what they build. The second wave is about control. The AccountingWEB article, published in association with Caseware, describes audit moving away from open experimentation towards governance, traceability and repeatable results. Bradley’s distinction is between an agent that works as a proof of concept and one that works consistently at scale.

Regulators are pushing the same way. The Financial Reporting Council’s March 2026 guidance on generative and agentic AI makes clear that responsibility for audit quality stays with firms and human auditors. Under ISQM (UK) 1, firms have to consider technology within their quality systems, which brings vendor selection, testing and documentation into the same conversation as productivity.

The wider finance world is reaching the same conclusion. Writing in Accountancy Age, iplicit CFO Rob Steele cited Workday and Hanover Research findings that nearly 40% of the time saved by general-purpose AI tools is lost to rework: correcting errors, rewriting text and checking outputs. His piece is a partnered article and his argument suits his product, but the point holds. Plausible output that has to be checked isn’t the productivity gain it first appears to be.

Most accounting firms don’t carry out audits, so it’s tempting to file this under someone else’s problem. It isn’t, for two reasons. The governance questions apply to any firm using AI on client work, and the same logic of evidence now shapes how firms get found.

What independent accountants need to know

Your clients will ask about your AI use. Bradley suggests firms should ask what problem a tool solves, how they’ll know it’s working and what could go wrong. He also recommends looking at where data is held and how updates are tested. Those are good questions for any tool you use, and a client or prospect could soon ask you the same things. Have an answer ready.

Specialist doesn’t automatically mean safe. The article makes the point that software isn’t trustworthy just because it sits inside a professional workflow. Firms still need to know how outputs are supported and whether results stay consistent after an update.

AI search behaves more like an auditor than an advertiser. When a prospect asks an assistant for a good accountant for a small construction business in their town, the tool can only draw on what it can find and cross-check. It looks for consistent facts across your website, your Google profile, directories and reviews. A firm making vague claims gives it little to work with. A firm with named specialists, clear services and specific examples gives it evidence.

Generic AI content makes you less visible. Many firms have responded to AI by using it to produce more blog posts. If those posts could have come from any firm, in any town, they prove nothing about yours. Plausible is exactly what AI tools can already generate for themselves.

What forward-thinking practices are already doing

Apply audit’s new discipline to your own firm, and a practical to-do list emerges.

They’ve moved from everyone experimenting to a few defined uses. Instead of leaving staff to their own devices, they’ve chosen two or three bottlenecks, such as drafting client emails, summarising bank data or preparing meeting notes, and written down who checks the output. A one-page AI policy covering approved tools, client data and human review is achievable for any practice.

They’re publishing that policy. A short, plain statement on the website about how the firm uses AI, what stays with a qualified person and how client data is protected reassures prospects. It also gives AI search tools a clear, citable fact about the firm that competitors may not have.

They’re auditing their own digital footprint. Firm name, address, services, specialisms and staff qualifications are checked for consistency across the website, Google profile, LinkedIn, directories and professional body listings. Conflicting information weakens the evidence, just as it would in a file review.

They’re replacing vague claims with proof. “We’re proactive and friendly” becomes “we look after 60 trades businesses across Shropshire, and our tax team is led by a named chartered tax adviser”. Those figures need to be real, specific and kept up to date. Case studies, dated articles, author bios and detailed reviews all add to the record.

They’re writing to answer questions. Pages are structured around the questions clients ask, with direct answers at the top, clear headings and named sources. This helps human readers, and it’s the core of practical GEO (Generative Engine Optimisation).

How this connects to growth

AI SEO isn’t a separate discipline from good marketing. It’s what good marketing looks like when the first reader is a machine deciding whether to recommend you. The firms that benefit will be those whose expertise can be verified.

Foundations still matter. Strong local SEO, a well-maintained Google My Business for accountants profile and a steady flow of genuine reviews give AI tools the raw material to work with. Add specific, expert content and the firm becomes easier to recommend, whether the search starts on Google, in ChatGPT or in a Copilot window on someone’s desktop.

There’s a commercial upside for smaller firms as well. Independent accounting firms can’t outspend large groups on technology, but they can often show deeper local and sector knowledge. An independent practice with real experience in farming, dentistry or construction in one region has something an AI assistant can match to a specific question. A national brand’s generic service page may not be as useful for that query.

This all feeds lead generation for accountants. Prospects who arrive after an AI recommendation have often already compared options, so the enquiry is warmer. Your site then needs to confirm what they’ve been told, with the same facts, the same specialisms and an easy next step.

Evidence wins

Audit has worked out that the question isn’t how much AI can produce, but whether a person can stand behind the result. Practice owners should ask the same question about their marketing. If a prospect, a regulator or an AI assistant checked what your firm says about itself, would the evidence hold up? A firm whose answer is yes will be recommended more often. A firm relying on plausible-sounding copy is taking a risk it can’t see.

Independent practices that want to be visible and verifiable in AI search are finding that it takes more marketing expertise than most firms keep in-house. The CharterGroup Alliance gives forward-thinking practices that expertise across AI SEO, GEO, local SEO and website management, along with a network of peers working through the same changes. If you want your firm’s reputation to show up where clients now look, take a look at CharterGroup membership.

Published by the CharterGroup team