If someone asks ChatGPT or Google's AI Overview "who's a good accountant for a small business in Leeds," the tool looks for pages and listings that help answer it. If your firm never wrote that answer, it's harder for the tool to name you, however good your accounting is.
1. Why AI recommendations matter now
Search behavior has split. People still type into Google, but a growing share now ask ChatGPT, Perplexity, or Google's AI Overview directly, and get a short list of names back instead of ten blue links. For an accounting firm, that's a new front door, and most firms haven't noticed it opened.
The firms getting named aren't necessarily the biggest or the longest established. They're the ones whose website gives a clear, direct answer to the exact question a prospect is asking. AI tools don't weigh reputation the way a human referral does. They weigh clarity. A page that says "we handle VAT returns for contractors in under 48 hours" beats a vague "comprehensive accounting solutions" every time, because the AI tool can match specific language to a specific query.
This matters more for smaller and mid-sized firms than for the big names, because AI recommendations are one of the few channels where a well-written page can outrank a firm ten times your size. The playing field is flatter here than it's ever been on traditional Google rankings.
2. What AI tools actually read
AI tools read your website the same way a very literal, very fast new employee would. They pull text, not design. A beautiful homepage with a hero video and three words of copy gives an AI tool almost nothing to work with. A page with full sentences explaining what you do, who you do it for, and what it costs gives it plenty.
These tools also read structured data behind the scenes, which is a separate layer from the text a visitor sees. Getting the right schema markup on your site tells an AI tool exactly what kind of business you are, what services you offer, and what questions you've already answered, without it having to guess from the design. Most accounting firm websites skip this entirely, which hands an advantage to any competitor who hasn't.
Review sites, directory listings, and local business profiles also feed into what AI tools know about a firm. If your Google Business Profile says one thing and your website says another, the tool has to pick one, and it often picks neither.
3. Write content that answers the question being asked
The firms that get recommended write pages that match real questions word for word. Someone doesn't search "chartered accountancy services." They search "accountant for a limited company director in Manchester" or "who does tax returns for landlords." A service page that mirrors that exact phrasing gets pulled into the answer. A service page full of generic language about "trusted expertise" gets skipped.
This means every service page needs its own clear answer: what the service is, who it's for, roughly what it costs, and how long it takes. Vague promises read as filler to both humans and AI tools. Specific answers read as substance.
A clear FAQ section does a lot of this work by itself. Questions and answers are the exact format AI tools are built to extract and quote back to a user. A firm with ten well-written FAQ entries covering real client questions is handing an AI tool ready-made answers it can lift directly. A firm without one is making the tool guess, and tools that have to guess usually move on to a competitor instead.
4. Structure your site so machines can parse it
Clear headings matter more than they used to. An H2 that reads "How much does a small business accountant cost" gives an AI tool a direct match to a direct question. A page with no headings, or headings that are just decorative ("Our approach"), gives it nothing to grab onto.
Short paragraphs help too. AI tools tend to extract two or three sentences at a time, not full paragraphs of dense copy. If your best point is buried in the middle of a 200-word paragraph, it's far less likely to get pulled out and quoted than if it sits in its own short, direct sentence near a relevant heading.
Page speed and basic technical health still matter here, because a slow or broken page sometimes doesn't get crawled properly at all. If your site has structural problems, whatever AI tools might have recommended never reaches them in the first place.
5. Build the trust signals AI tools look for
AI tools lean on signals of legitimacy before they'll name a business with confidence. Professional accreditations, named team members, client testimonials with real names, and years in business all feed into whether a tool treats your firm as a credible answer or skips past it. Trust signals that used to exist mainly to reassure a human visitor now do double duty, feeding the same confidence into the machine reading your site.
A firm that hides its team behind generic stock photos and a "Contact Us" form is invisible in a different way than it used to be. It's not just that visitors trust it less. It's that AI tools have less to point to when deciding who to recommend.
6. Keep your information consistent everywhere
AI tools cross-reference. If your website says you're based in Bristol, your Google Business Profile says Bath, and an old directory listing says you closed in 2019, the tool either picks the most recent clean source or avoids naming you at all to be safe. Consistency across your website, Google Business Profile, LinkedIn, and any directory listings is no longer just good practice for local SEO. It's what lets an AI tool trust the facts enough to repeat them.
This is also where a lot of firms quietly leak credibility. An outdated "our team" page, a phone number that's been disconnected for a year, a services list that hasn't been touched since before a merger. None of that is dramatic on its own, but together it tells both humans and machines that the information in front of them might be stale, and stale information doesn't get recommended.
Conclusion
You don't need a separate AI plan. You need clear, current pages that a client can understand too. Today, write one sentence for your homepage that says who you help, where you are and what you do. Have a partner check it for accuracy and professional rules. Then ask ChatGPT and Google who they'd recommend, and check the same question monthly.
FAQ
Do AI tools like ChatGPT actually recommend specific accounting firms?
They can. Tools like ChatGPT, Perplexity and Google's AI Overview sometimes name specific firms, drawing on firm websites, directories and review platforms. A growing share of people now ask these tools directly and get a short list of names instead of ten blue links. The firms named are often the ones whose site answers the exact question asked, whatever their size, but nobody can promise a mention.
Does my accounting firm need schema markup to be recommended by AI?
No. You don't need special schema for AI features, and schema isn't a direct ranking boost. It can help describe your firm clearly, but the words on your pages matter more. Start by writing clear answers to the questions clients ask, structuring pages so they are easy to read and keeping your details consistent everywhere.
Will a nicer-looking website help me get recommended by AI?
Design helps human visitors, but it won't do much for AI tools, which rely on text and page structure. Clear, specific answers to real client questions are the better place to start. If someone asks for "a good accountant for a small business in Leeds" and your firm never wrote that answer, the tool has less to work with. A clear FAQ is the format these tools are built to read.
How long does it take to start showing up in AI recommendations?
There is no fixed timeline, and nobody can promise one. Rewrite your service pages and FAQs with clear answers, then check the same questions each month to see what changes. A good first step is one homepage sentence that says who you help, where you are and what you do. Have a partner check it for accuracy before publishing.
Is this different from normal SEO?
It overlaps a lot with SEO. SEO aims to rank in a list of links, while writing for AI tools focuses on short, specific answers a tool can understand and summarise. You don't need a separate AI plan. Clear, current pages that a client can understand do both jobs.









