If you run an engineering company, there's a decent chance you've got some cracking case studies. You've solved complex problems for demanding clients, delivered projects on time and on budget, and you've got the photos and testimonials to prove it. The trouble is, the way most of those case studies are written and published means that AI search tools (we're talking ChatGPT, Perplexity, Google's AI Overviews and the like) can't really make sense of them, which means they're not going to recommend you when someone asks "who's good at subsea inspection in the North Sea?" or "which engineering companies specialise in modular fabrication?"
We're Red Evolution, a B2B web design and digital marketing agency in Aberdeen, and we work with a lot of engineering companies. We've been watching closely as AI search has started to change how buyers find suppliers, and if you've read our guide to AI search and B2B, you'll know we think this shift is a big deal. So this post is about one specific thing you can do right now: making your case studies AI friendly so that when someone asks an AI tool for a recommendation, your company is in the conversation.
What do we mean by "AI friendly"?
Let's get the jargon out of the way first. When we say "AI friendly case studies" we're talking about case studies that are structured, written, and published in a way that AI tools can easily read, understand, and use to answer questions. That's it. It's not complicated in principle, but most engineering companies get it wrong because they're still writing case studies the way they did in 2015, and AI tools need something a bit different from what a human reader needs.
Here's the thing. When a procurement manager in Houston asks ChatGPT "which UK companies provide pipeline inspection services?", the AI doesn't go and browse your website the way a person would. It pulls from a massive pool of text it's already ingested, or (in the case of tools with live search like Perplexity) it crawls the web and tries to make sense of what it finds. Either way, it needs your content to be clear, specific, and structured in a way it can parse. A beautifully designed PDF case study with lovely photos and vague marketing copy about "delivering excellence" is, unfortunately, almost useless to it.
Why most engineering case studies don't work for AI
We've looked at hundreds of engineering company websites over the years, and there are a few patterns we see again and again that make case studies effectively invisible to AI search tools.
The first is the PDF problem. A huge number of engineering companies publish their case studies as downloadable PDFs. Sometimes they're gorgeous, properly designed documents, but AI tools struggle with PDFs. They can't always read them, they can't index them reliably, and even when they can extract the text, the formatting often comes out garbled. If your case studies only exist as PDFs, AI tools are probably ignoring them entirely.
The second is vague language. Engineering companies are, understandably, sometimes cautious about what they share. Client confidentiality, commercial sensitivity, that sort of thing. But the result is case studies full of phrases like "a major international operator" and "a challenging offshore environment" and "we delivered the project to the client's satisfaction." That's not enough for an AI tool to work with. It needs specifics: what sector, what service, what geography, what problem, what outcome.
The third is missing structure. A lot of case studies read like short essays, a couple of paragraphs with no headings, no clear separation between the problem, the solution, and the result. Human readers can usually figure out what's what, but AI tools are much better at parsing content when it's clearly organised with headings and a logical flow.
How to make your case studies AI friendly
Right, so what do you actually need to do? We'll walk through this step by step, and honestly, none of it is rocket science (forgive the engineering pun). It's mostly about being more deliberate with how you write and publish.
1. Put them on your website as proper web pages
This is the single most important thing. Each case study needs to be its own page on your website, with a proper URL, written in HTML that search engines and AI tools can crawl. Not a PDF. Not a carousel. Not hidden behind a form that makes people hand over their email address before they can read it. A proper, crawlable web page.
You can still have a PDF version as a download if you want (some people still like to print things out or email them to colleagues), but the web page version is the one that AI will read.
2. Use a consistent, clear structure
Every case study should follow the same basic format, and you should use proper headings to separate each section. Something like this works well:
- Client and sector (who was this for, and what industry are they in?)
- The challenge (what was the problem or requirement?)
- What we did (your solution, in specific terms)
- The results (measurable outcomes wherever possible)
- Services used (a simple list of the capabilities you deployed)
That last one is easy to overlook, but it's really useful for AI. When someone asks "who provides NDT services in Aberdeen?", the AI is looking for pages that explicitly mention NDT as a service. If your case study describes the work you did but never actually names the service category, you're missing out.
3. Be specific, even when it feels uncomfortable
This is where a lot of engineering companies struggle, and we understand why. Clients sometimes don't want to be named, contracts have confidentiality clauses, and there's a natural instinct to keep things vague. But you need to push back on that instinct as much as you reasonably can, because specificity is what makes your content useful to AI tools.
If you can name the client, name them. If you can't, at least be specific about the sector and the type of company ("a North Sea oil and gas operator with a fleet of 12 production platforms" is infinitely better than "a major energy company"). Include real numbers wherever possible: project duration, cost savings, efficiency improvements, anything quantifiable. And name the geographic location, because a surprising amount of AI search is location-specific.
If you're wondering how to write this kind of content well, our post on content marketing for engineering companies goes into more detail about finding the right balance between technical depth and readability.
4. Write in plain English
AI tools are good at understanding natural language, which means the best thing you can do is write the way you'd explain the project to someone at a conference. Clear, direct, no unnecessary jargon. If you need to use technical terms (and in engineering, you obviously will), that's fine, but make sure the surrounding context makes it clear what you're talking about.
We've written before about why plain English matters on your website, and it's even more relevant now that AI tools are reading your content. They're trained on natural language, so writing naturally is, quite literally, the most AI friendly thing you can do.
5. Include the questions your buyers actually ask
This is a trick that a lot of companies miss. Think about the questions that procurement managers, project engineers, and technical directors type into AI tools when they're looking for a supplier. Things like "who provides corrosion monitoring for offshore wind farms?" or "which companies do pressure vessel fabrication in Scotland?"
Now make sure your case studies contain those phrases naturally. Not stuffed in awkwardly, but woven into the text. "We were brought in to provide corrosion monitoring for a 1.2GW offshore wind farm in the North Sea" is a sentence that does two jobs at once: it tells the human reader what happened, and it gives the AI tool exactly the kind of language it needs to match your company to a relevant query.
If you're not sure what questions your buyers are asking, our guide to keyword research will help you figure that out. The same methodology works for identifying the queries people are putting into AI tools.
6. Add structured data if you can
This one's a bit more technical, so you might need your web developer for it. Structured data (sometimes called schema markup) is a way of labelling your content so that search engines and AI tools can understand it more easily. Think of it as putting name tags on different parts of your case study: "this is the client name", "this is the service we provided", "this is the location", "this is the outcome."
It's not strictly essential, but it helps, and if your competitors aren't doing it (most engineering companies aren't), it gives you an edge. Your web developer will know what to do if you ask them about adding schema to your case study pages.
A quick word about what not to do
We've seen a few engineering companies respond to the AI search trend by trying to game the system, stuffing case studies with keywords, writing content that reads like it was generated by ChatGPT (which, let's be honest, it probably was), or creating dozens of thin, low-quality case studies in the hope that more is better. None of this works. AI tools are getting better at spotting low-quality content, and more importantly, your human readers will spot it immediately and lose trust in you.
The goal is to give AI tools genuinely useful, well-structured information about what you do, so that when a relevant question comes up, your company is a credible answer. Trying to game it will backfire, and probably faster than you'd think. If you're interested in how this fits into a broader strategy, we've written a more detailed piece on AI search strategy for marketers that's worth a read.
Where to start
If you're sitting there looking at a website full of PDF case studies and vague project descriptions, don't panic. You don't need to redo everything overnight. Pick your five best projects (the ones you'd most like to win again), and rewrite those case studies first, following the structure we've outlined above. Get them published as proper web pages, make sure they're specific and well-structured, and then see what happens. You can always add more later.
And if you're not sure where your website stands right now, or you'd like a hand restructuring your case studies to work better for both AI and human readers, we're happy to have a chat. No hard sell, no obligation, just a 15-minute conversation to see if we can point you in the right direction. You can get in touch here.
