There are managers asking ChatGPT to write the copy for their website. The result is grammatically correct, structurally reasonable and completely generic in content. It could be any company in the sector. And for a B2B business that sells trust before price, that is exactly the worst possible starting point.
Artificial intelligence doesn’t know who you are. It doesn’t know which companies buy from you, what objections your salespeople face on every call, or why one of your clients chose you over the cheaper alternative. That isn’t a minor limitation — it’s the core of the problem when AI is used to build corporate digital presence without prior strategic judgement.
Why AI can’t replace business context
Generative AI models work well when the task is generic: drafting a formal email, summarising a document, translating a paragraph. They work badly when the task requires specific knowledge nobody gave them.
Your company has history, a competitive position, a type of client that recognises value in concrete things, and a proposition resting on reasons written down in no public document. AI, by definition, works with what exists in its training data plus whatever you dictate at the moment of the query. If you don’t tell it that your industrial services company has spent 18 years working exclusively with the mining sector, it won’t infer it. And if you do tell it, it uses that during the conversation and forgets it in the next one.
That forgetting has real consequences for a corporate website. A B2B site isn’t a handsome brochure: it’s the place where a procurement manager who has never met you decides whether it’s worth booking a meeting. For the site to do that job, it needs to communicate something specific with surgical precision. AI, without context, produces the opposite: plausible but empty copy that sounds fine and says nothing that differentiates.
The underlying problem: AI generates content, not strategic judgement
There’s a frequent confusion among companies approaching AI for their B2B website: they believe the problem they’re solving is content production, when in fact the problem is strategic judgement.
The content of an effective B2B corporate site doesn’t come from writing well. It comes from answering hard questions before a single line is written: who exactly is this site speaking to? What main objection does it need to resolve before that visitor calls anyone? What proof of competence does the managing director of an 80-person company need to see to decide this isn’t a generic supplier?
Those questions require diagnosis, not generation. And that’s where it breaks.
When a company asks an AI tool to “write the services page”, the tool generates text based on statistical patterns of how services pages look across the internet. The result is an average. And an average doesn’t convert in contexts where the buying decision involves multiple people, significant sums and long sales cycles.
An example that makes it concrete
Imagine two project management consultancies. Both ask an AI to generate the text for their main page. Both receive something similar: “We help companies optimise their processes and achieve their objectives with proven methodologies.”
The first works exclusively with mid-sized construction firms on public infrastructure projects. Their clients have immovable legal deadlines and the cost of a delay is measured in contractual penalties. The second works with technology startups that need to scale teams fast.
The problem each of them solves isn’t the same. Their client isn’t the same. The proof of competence each needs to show isn’t the same. But their sites will say the same thing, because the AI had no context to tell them apart.
How to assess whether your site has the lost-context problem
Before hiring anyone or investing in anything, there are concrete signals that indicate whether your site is operating as a generic digital asset rather than a specific business tool.
Quick diagnostic checklist — apply it yourself:
-
Read your homepage’s main copy out loud. Could it be a direct competitor’s text without changing a single word? If yes, your company’s context isn’t there.
-
Identify who your site explicitly speaks to. Does it name the type of company, the decision-maker’s role, or the sector you serve? If the site speaks to “everyone”, it speaks to nobody who is going to buy.
-
Look for specific proof of competence: cases with measurable results, named clients (where the agreement allows), certifications or proprietary methodologies. AI generates generic assertions — real context is demonstrated with specific evidence.
-
Check whether the site explains the problem you solve, not just what you do. “Industrial maintenance services” describes a category. “Reducing unplanned stoppages in continuous production plants” describes a problem someone is willing to pay to solve.
-
Assess the information architecture: does the order of sections match the mental process of a client arriving cold? Or does it match how the company is organised internally?
If you failed three or more of these, the problem isn’t aesthetic or technical in the first instance — it’s strategic.
How to use AI without losing the context your company needs
AI is useful. The PTW team isn’t arguing against using it — it’s arguing against using it without a base. The difference between a site that converts and one that doesn’t usually lies in the order of decisions, not in the tools.
The right principle is this: context first, production second.
Step 1: Document the context before asking for any copy
Before opening any AI tool, build a context document that explicitly answers:
- What type of company do we sell to? (size, sector, decision-maker’s role)
- What is the concrete problem we solve, in the terms the client would use, not the supplier?
- What proof of competence do we have that a competitor can’t copy?
- Why did our last important client choose us?
- What objection always appears before anyone signs?
With that document as the base, AI can help produce copy that isn’t generic, because the specific context is already in the prompt. Without it, AI produces averages.
Step 2: Use AI to optimise, not to define
AI is more useful as a refinement tool than a starting point. If you’re clear on what you need to say and to whom, you can ask it to improve the clarity of a paragraph, propose variations on a headline, or identify whether a section is redundant. That’s different from asking it to “write the company’s value proposition” with no other input.
Step 3: Validate with someone who doesn’t know your company
Show your homepage copy to someone who doesn’t work with you and resembles your ideal client. Ask them to tell you, after 30 seconds of reading, what your company does, who it serves and why they should call you rather than look elsewhere. If they can’t answer all three, the site has a context problem regardless of how the copy was generated.
The technical base matters too: technical SEO and AI aren’t the same problem
There’s another angle that appears in any discussion of AI and B2B websites: the impact on organic ranking. Google has stated it doesn’t penalise AI-generated content per se — it penalises low-quality content, whatever its origin. But there is an indirect consequence that does matter.
Sites that generate content en masse with AI and no editorial selection tend to produce pages with low specificity, thin topical authority and high semantic similarity to each other. That affects technical SEO, because Google rewards sites demonstrating depth and domain authority on a topic, not a volume of pages saying the same thing in different words.
For a B2B company that wants to appear when a procurement manager searches for a specific supplier, the site’s technical base and the quality of its content are inseparable. A technical SEO diagnostic assessing structure, crawlability and search intent is more relevant today than it was three years ago, precisely because the volume of mediocre content online has grown exponentially with generative AI.
Before using AI on your site, there’s a more logical step
If after reading this you recognise that your site speaks generically, that it isn’t built for your ideal client’s mental process, or that it was produced with tools that had no context about your company, the question isn’t which tool to use next. The question is what the site needs to say before deciding how to produce it.
That is exactly the work PTW’s strategic diagnostic does: identifying what’s failing in the current digital presence, why it fails, and what has to change for the site to work as a business asset. It isn’t a design audit. It’s an analysis of whether the site is built to convert the type of company you need as a client.