Claire Taylor of TU Marketing has been running her B2B agency for 19 years, works with founders £1m to £10m, and writes for Search Engine Land.

I had a chat with a founder recently who had just watched a competitor’s name appear in a ChatGPT answer where their own did not.

“Do we need an enterprise budget to compete on this?”

The short answer is no. The longer answer is that most small B2B businesses are already sitting on the raw material they need. They just have not organised it in a way a machine can read.

The assumption to knock down first

There is a widely held belief that AI visibility is a big-brand game. That ChatGPT and Perplexity and Google’s AI Overviews recommend businesses with enterprise SEO budgets, PR agencies, and a research team feeding them citations.

They do not. They recommend businesses they can identify with confidence.

Confidence here means clarity. Size has nothing to do with it. An AI assistant asked to recommend a supplier is looking for the business it can describe accurately in one sentence: who you are, what you do, who you do it for, where you serve. Big companies often fail this test because their websites are written by committee and their positioning has been sanded smooth by legal. Small businesses often fail it too, but for the opposite reason: they never wrote the sentence down at all.

The good news is that writing it down does not require a budget. It requires a decision.

Where small businesses actually have the edge

Enterprises have money. Small B2B businesses have three things enterprises struggle to buy at any price.

Specificity. A £2m turnover business that sells only to owner-managed manufacturers in the Midlands can say so, in those words, on the homepage. A £2bn business that sells to everyone cannot, so it says nothing.

A founder who can talk. Google’s AI Overviews, ChatGPT, Perplexity: all of them reward businesses that show up in interviews, podcasts, guest articles, comment threads. A founder who can spend two hours a week doing that, consistently, will outrun a marketing department that needs a sign-off chain to publish a LinkedIn post.

Real client work you can describe honestly. A named client with a specific result in a specific sector is worth more to an AI assistant than a page full of Fortune 500 logos with no story attached. Small businesses have the client relationships to write the honest version. They rarely bother.

None of that costs money. All of it costs deliberateness, which is harder.

Four things that actually work

Say what you do in plain English, on the pages you already have.

AI systems extract facts more reliably from short declarative sentences than from marketing prose. One of our clients is a contract electronics manufacturer, and the most valuable line on their site is one sentence: “We are a contract electronics manufacturer offering free NRE (the upfront engineering set-up costs) for UK SMEs.” Machine-readable and human-readable. A paragraph about your passion for excellence is neither. The most valuable half-day you can spend this month is rewriting your homepage, services page and about page in the voice of someone answering a question at a networking event.

Publish answers to the questions your buyers actually ask.

Not thought leadership. Not trends pieces. Actual questions, answered plainly, on your site. That same manufacturer gets asked “how much does contract electronics manufacturing cost for a low-volume run?” A page that answers that question directly is a page an AI will quote. A page that ends “it depends, get in touch” is not.

Make your entity signals match everywhere.

Same business name, same service description, same location on your website, your Google Business Profile, LinkedIn, industry directories, and every profile you have. AI systems build confidence by cross-referencing. Inconsistency reads as uncertainty, and uncertain businesses do not get recommended. This is a free afternoon of copy-paste that pays back for years.

Get quoted somewhere other than your own site.

One guest article on a trade publication your buyers read is worth more to an AI assistant than ten posts on your own blog. The reason is simple: your own site says nice things about you, and the model knows that. A third party saying you know what you are talking about is a different class of signal. Aim for one placement a quarter, minimum.

What is a waste of time

For most small B2B businesses, the following are noise: buying expensive “AI SEO” tools that produce reports nobody reads, obsessing over which specific ChatGPT prompt returns your name, and paying for links from directories nobody has heard of. All three cost money and none of them will show up anywhere you can measure.

What to expect, and over what timeframe

The fastest we have seen this work is a client three months in, with ChatGPT now showing up in their analytics as a traffic source and, more to the point, converting into sales. That is the exception, and it came from doing everything above at once. Six months to see the first movement is more typical. Twelve months to see it stick. Longer if you are in a competitive sector.

Anyone quoting weeks is either selling you something or wrong.

The businesses winning at this right now are not the ones with the biggest budgets. They are the ones who started twelve months ago and did not stop. The gap between them and everyone else is going to widen, not narrow, and it will not be closed by a bigger cheque.

It will be closed by writing your positioning down, saying it clearly, and repeating it, in public, for a year.

Start this week.

Claire Taylor is the founder of TU Marketing, a B2B digital marketing agency working with founder-led businesses turning over £1m to £10m. She is a contributor at Search Engine Land.