Paul Evans wants to know how ChatGPT describes your business when you're not in the room
The website spent thirty years learning to sell. Then buyers stopped reading it and started asking a machine. Paul Evans, founder of V2RSION, has a plan for what comes next.
There is a sales conversation happening about your company right now, and you are not invited. It is short, it is confident, and it ends in a shortlist. On one side is a prospective buyer with a problem and a budget. On the other is not a salesperson, not an analyst, not a peer in a WhatsApp group. It is ChatGPT. Paul Evans has spent the last stretch of his career trying to get companies to take that conversation seriously.
Evans is the founder and Chief Positioning Engineer at V2RSION, a London studio he describes as the first positioning engineering practice built for the intelligence era. Before that he ran media at Vodafone as its Global Head of Media, worked on go-to-market strategy inside Microsoft, led brand work at Emirates, and pushed marketing innovation at scale at Nike. Twenty-odd years of watching how buyers decide, and a growing suspicion that the rules had quietly changed underneath everyone.
His argument starts with an uncomfortable observation. The B2B website, that expensive, endlessly A/B tested shrine to conversion, is losing its job. Buyers may never visit it. When they do, it is often to confirm a choice they already made somewhere else. The persuasion now happens upstream, in a chat window, before a human has clicked anything.
"If AI can't understand what makes your business different, it can't recommend you. And if it can't recommend you, you may never make the shortlist."
The shiftWhere the shortlist actually gets made
For a generation, marketing operated on a comforting premise: content sells and converts. You published, buyers read, some of them bought. Search engine optimisation was the tollbooth, and everyone learned to pay it. Evans thinks that premise is now only half true, and the wrong half.
Buyers are asking ChatGPT for recommendations. They are using Gemini to compress a week of research into an afternoon. They are lining up providers side by side in Perplexity and asking, plainly, which one is better and why. By the time they reach your site, the field has already been narrowed by a machine that never filled in your lead form.
Where B2B buyers now form a shortlist
The implication is blunt. Prospect influence now flows through large language models. And an LLM cannot recommend what it cannot recognise. If the model reads your site, your competitors' sites, a dozen review pages and a stack of PDFs and comes away describing all of you in the same beige sentence, you have not been beaten. You have been blurred.
The diagnosticSix questions, one honest afternoon
Evans does not open with a framework. He opens with a dare. Go and ask ChatGPT to describe your business. Then ask it to describe your three closest competitors. Read the four answers back to back. If you cannot tell which one is you, neither can your buyer's AI. From there he offers six questions worth sitting with.
The readiness test
Does AI describe your business differently from your competitors? If every answer sounds the same, your positioning lacks distinction.
Do you appear when buyers ask AI for recommendations? Absent from the conversation means absent from consideration.
Is your differentiation obvious without a human to explain it? AI can only surface what it can clearly recognise.
Is your message consistent everywhere - site, LinkedIn, customer stories, sales calls? Consistency makes you easier to understand and trust.
Does your team still have to explain what makes you valuable? If every call opens with a long preamble, AI won't bridge that gap.
Is your expertise easy to find? Original thinking, research, frameworks, case studies - the more evidence, the more authority.
Read them again and a pattern surfaces. Not one of these is about publishing more. They are about being legible. Consistent. Distinct. The questions assume the problem is rarely a shortage of words and almost always a shortage of clarity.
"This isn't an SEO conversation. It's a positioning conversation."
The methodPositioning as engineering, not decoration
The job title gives away the worldview. Evans calls himself a positioning engineer, and he means the noun. Positioning, in his telling, is not a tagline you commission once and frame in reception. It is infrastructure. Something you build, instrument, and keep tuning as the market moves under you.
V2RSION packages this into what Evans calls Positioning Engineering and a 5D Positioning Framework, driven by a diagnostic he has named Positioning Intelligence, or PIQ. Strip away the acronyms and the idea is stubbornly practical: combine AI, market research and competitive analysis to define what makes a company genuinely preferable, then keep that definition current as rivals shift and language drifts. Static messaging, he argues, is no longer an option, because the audience reading it is no longer static either.
There is a nice irony in a former media chief deciding that the answer to an AI problem is not more media. Evans is careful about it. The companies that win AI-driven discovery, he says, will not be the ones who flood the internet with content. They will be the ones with positioning so clear that both a person and a model can repeat it back without help.
The winners won't simply produce more content. They'll build positioning that people and AI can both recognise.
The personRecords, frameworks, and a habit of naming things
Evans has a taste for building categories rather than joining them. He coined a job title before the market had a box for it. He has a Cambridge grounding in sustainability leadership, a Campaign 40 Over 40 award, a place in the Media Leader Future 100 Club, and - the detail that tends to derail an otherwise sober CV - a Guinness World Record. He is also a vocal advocate for diversity, equity and inclusion, and for neurodiversity in particular, which sits comfortably with a man who keeps insisting that being different is the whole point.
That thread runs through the work. In a market drifting toward sameness, where every SaaS homepage promises to be intuitive, scalable and end-to-end, Evans treats distinction as an engineering requirement rather than a creative flourish. The model is not going to invent your difference for you. It can only report the difference you have made findable.
Which brings the argument back to that conversation you were not invited to. You cannot join it. You cannot sit beside the buyer and interrupt the model mid-sentence. What you can do is decide, in advance, what the machine has to work with. The evidence, the framing, the consistency, the proof. Feed it clarity and it can carry your case. Feed it fog and it will hand your buyer to whoever was easier to understand.
Evans ends most of his arguments the same way, with a question that is quietly ruthless: have you asked ChatGPT how it describes your business lately? Most executives have not. The ones who have tend to go quiet for a moment. Then they open a new tab.