How technical B2B businesses can win in AI search recommendations
- Kristy Dodd
- Jun 10
- 6 min read
Updated: Jul 11
A question I keep getting asked is some version of this: should I be worried about showing up in AI search? Am I missing out if I’m not appearing in AI overviews? It’s not a worry. It’s an opportunity. And for many technical and industrial businesses, it’s the first real chance at organic visibility they’ve had.
Most of the technical B2B businesses who reach out to KLD Marketing for help have been paying for Google ads, because organic search has always felt like a game stacked against them. AI search changes that equation. If your website speaks directly to the problem your buyers are having, in language that reflects how they actually think about it, you have a genuine chance of appearing ahead of a company ten times your size. Not because of a trick or a technical fix. Because you gave AI something useful to work with, and they didn't.
Industry articles on this topic, Google’s guidance, Semrush’s breakdown of their own approach, The Microsoft Framework, they all point to the same B2C SEO fundamentals. What I haven’t found is anything that tests these principles specifically for B2B technical industries. Mining, infrastructure, engineering, construction. So I conducted original research across five AI platforms, using real queries that buyers in these sectors actually type.
How I tested AI search recommendations across five platforms
The research involved multiple rounds of testing across AI platforms, with findings cross-referenced against published research on AI search behaviour. I ran two types of searches: a broad category and a specific problem query. What kept coming back was the same pattern.
The broad search was a straight industry category, the kind a buyer types when they already know what they need and are looking for who can supply it. The specific search described a real situation, the kind a buyer types when they know they have a problem but aren't yet sure what the solution looks like or who to call, something like "I work on a mine site in Qld and we have outdated manual processes for monitoring our blasting activities, they are time consuming help me find a better solution".
I expected the results to be similar but more refined for the specific query. What came back was a completely different set of companies, on every platform. Not a refinement. A different list entirely.
Broad category searches favour national suppliers
For the broad category search, large national suppliers appeared consistently across every platform I tested. Not because they had the best websites. And not just because they ranked first in Google. Because they have been talked about, referenced, and cited consistently across the web over a long period of time.
Trade media. Industry directories. Supplier partner pages. LinkedIn mentions. Case studies published elsewhere. Consistent presence across multiple source types, built up over years.
This is how an AI platform forms a confident recommendation. It draws on everything it has learned from across the web. The businesses that appear reliably in broad AI recommendations are the ones with the deepest footprint beyond their own website, not just the ones with a well-optimised homepage.
For broad category searches, having a good website is one input. How broadly you are talked about everywhere else is another, and for many platforms it carries more weight.
Problem-specific searches surface different businesses entirely
The specific problem query changed everything.
The major companies that dominated the broad search largely disappeared. In their place were specialist businesses whose websites spoke directly to the problem described in the query. Companies that explicitly addressed the compliance standard relevant to the work. Businesses that named the industries they serve, described what they actually do, and used language that matched how a buyer thinks about the problem.
Location handling added another layer. One platform asked clarifying questions before returning results. Once it had confirmed my location, it gave significant weight to local presence, returning hyper-local recommendations that no other platform surfaced. Those businesses appeared not just because their content was relevant, but because the platform prioritised proximity once location was confirmed. Without that location signal, they were invisible. Another platform used my location automatically, without asking, and returned local specialists ahead of national players. The majors didn't appear at all.
The pattern across both query types was consistent. Problem-focused content surfaces specialists. Location signals surface local businesses. Neither advantage belongs to the major players.
Four things that actually determine AI search recommendations
Across both query types and all platforms tested, four distinct factors emerged. They are not interchangeable. Each one does different work.
How broadly you are talked about beyond your own website. This is the primary driver for broad category searches. It is built slowly, over time, through trade media, industry directories, associations, partnerships, and LinkedIn presence. It cannot be shortcut by website work alone. Think of it as your reputation across the web, not just on your own pages. Start with the directories and associations your clients already trust, get listed, stay current, and make sure your name appears where your industry already looks.
How clearly your website communicates what you do. When an AI platform visits your page, it tries to extract usable information from it. One platform I tested attempted to build a comparison table from retrieved pages. Where pages clearly stated who the business serves, what they do, and where they operate, the extraction worked. But where pages were dense or vague, or buried key information in general capability statements, the information was mangled or missing entirely. How your page is written is not just a design choice. It determines whether AI can read and use it. Read your own homepage out loud and ask whether someone who had never heard of you could explain what you do and who you help within thirty seconds. If they couldn't, neither can AI. I also cover how to write a service page for AI recommendations in the article below:
How well your website performs in Google. Some platforms pull recommendations directly from live Google search results, which means the basics still matter: clean page structure, descriptive headings, relevant content, and fast load times. This includes the SEO fields your website expects you to fill in but often gets skipped, your page title, meta description, and image alt text. These aren't technical requirements. They're the plain language labels that tell both Google and AI tools what your page is actually about. If your website platform has an SEO section you have never opened, that is the first place to start.
Whether your local presence is set up properly. Platform behaviour around location varied more than anything else I tested. Some platforms identified location automatically. Others asked. Others assumed a major city when no location was confirmed. For regional businesses, this means the same search query can return completely different results depending on which platform a buyer uses and whether their location is detected. The businesses that showed up consistently for location-specific queries had one thing in common beyond relevant content: a strong local presence that gave platforms something to match against. If you serve a regional market, that presence needs to be findable, not just assumed.
What this means for your content
The most important shift across all four of these is how you frame your content. The businesses that showed up in problem-specific searches weren't just well-optimised. Their pages spoke directly to the situation a buyer was in when they typed the query.
A capability statement says: we offer real-time vibration monitoring for your site. A problem statement says: we help blast crews demonstrate compliance with vibration limits in real time, so a neighbouring property complaint or a regulator query never catches you off guard.
Keywords in capability statements still matter and traditional SEO hasn't gone away. The shift is in how you lead. Start with the problem, then support it with the capability detail. Case studies are the most powerful way to do this. A well-written case study names the client, describes the problem they faced, and states what changed. Most technical businesses have that story in their heads or buried in proposals. Getting it onto the website as a published post is one of the highest-value changes you can make.
Where regional and specialist businesses actually win
Trying to outrank a major national supplier in a broad AI category search is the same problem as trying to rank number one in Google against a company with ten times your history and twenty times your budget. It is not where you win.
B2B and technical businesses can win AI search recommendations by being specific about the problems you solve and the clients you help. The majors dominate broad searches because they have broad presence. But broad presence does not make them the best answer to a specific problem. A buyer who knows what they need, who describes their situation in a search query, is not looking for the biggest name. They are looking for the most relevant answer.
That's not a new idea. It's the same thing that has always been true of good marketing. The businesses that stop trying to sell products or win jobs, and start trying to help people solve problems, are the ones that get found.
If you want to understand what AI is actually looking for when it visits your website, I've covered that in detail in the article below.
KLD Marketing helps businesses in mining, infrastructure, engineering, and construction build web content that works for both human buyers and AI search. If you’re not sure whether your website is doing that job, send me an email. I’m happy to take a look.
Comments