Article

Positioning as part of your AI optimization strategy

AI doesn't spot the best company. It recommends the ones it finds content about online, so AI visibility starts with positioning and traditional search.

Jussi Viljanen in a dark blazer with arms crossed against a grey background, with blue and green rays behind him.

AI has no magic way of surfacing the world’s best brands and companies. It surfaces and talks about the brands it finds material about online, and that is nowhere near in line with the real world. That is why an AI visibility strategy should be built on both positioning and search engine optimization.

Everyone wants to show up in AI these days, and companies dream of AI recommending them in particular. With good reason. The use of AI tools is growing, their conversion rates are high, and if your users and customers are there, you should be too.

A common misconception, though, is that AI somehow automatically recognises a good company: its long track record, its high-quality product or, say, its customer service. It doesn’t, and language models were never built to.

Language models are not brand-recognition tools. They are built to understand content: what is on the internet, or what they were fed in training.

If you prompt for a service or product, ChatGPT, Gemini or Claude won’t set out to evaluate every possible company. It simply can’t. For now, the most cost-effective and highest-quality way to find and compare companies is to use traditional search engines such as Google and Bing as the source.

When thinking about strategies and actions to increase visibility, these technical fundamentals come first. Only then does it make sense to think about how to actually earn that visibility and how to position your own company.

A prompt is not a search term

When you ask AI for products, services, recommendations, comparisons, prices or anything else more subjective, it almost always relies on web search. AI uses traditional search engines just like we humans do – only far more efficiently, through what is known as query fan-out.

AI breaks the prompt down into several search terms, runs them simultaneously in the background on traditional search engines, reads through the content of the results and synthesises its answer from them. The prompt gets its answer, but the mechanism underneath relies on traditional search engines.

The search terms AI generates, their length, style and emphasis can vary between tools, and from time to time within the same tool. Most of the time, though, they sound fairly conventional: terms of a few words, the kind we are used to typing into Google.

There is no need to make this more complicated than it is, as long as you understand the basic nature of query fan-out. The context contained in a prompt shapes the fan-out and the search terms in it.

Context decides what content you create and where you show up

Every search has a context: a frame of reference, a connection or something else that defines what the search means. That’s the fancy description.

In plain terms: if you search for the price of a service, the context is price. If you search for reviews, the context is reviews. If you search for a comparison, the context is comparison. If you search for the best, the context is “best”, and so on.

These contexts make their way into the search terms AI uses. If a prompt asks about price, among other things, I’d argue that the fan-out includes a search term with price-related words such as “price”, “prices” or “pricing” every single time.

If you want to rank in search engines for a price-related search term, you need a search-optimised page for it – in this case, the company’s pricing page. So the context of the search dictates what content to create, depending on what your own customers or target audience end up asking AI.

To understand which contexts its customers deal with and represent, a company first has to think about its own place in the market: its positioning.

Positioning defines your place in the market

By positioning I mean defining a company’s own place in the market and in the customer’s mind: who you are, what you do, for whom, for what need, and why you in particular.

Most companies surely have some kind of conscious positioning. For many, though, it is incomplete, far too broad or vague.

A broader positioning does suit some, if you sell everything to everyone like Amazon, but that applies to very few companies. Most businesses focus on a narrower group and target audience.

The driver: changing search behaviour

In AI optimization, positioning needs to be sharper and more precise. Almost the more precise, the better. Granted, this has always been part of in-depth search engine optimization too.

The need for even sharper positioning comes from changing search behaviour (article in Finnish). Where Google searches are around one to four words long, people use longer queries in AI chat interfaces and hold conversations, sometimes for days on a single topic.

Even though AI often runs its information retrieval with short search terms behind the conversation, the searcher – your potential customer – now uses clearly more precise, detailed and in-depth specifications, which flow into the query fan-out.

How to build a content strategy from your positioning

Once your positioning is clearer, it becomes much easier to map out your content strategy.

Take a simple, practical example: a SaaS company with its own software product. Its positioning work has clearly identified three industries where the software fits best and where marketing and sales make the most sense: construction, the chemical industry and the food industry.

These industries already give you an easy content plan for three landing pages:

  • “X software for construction companies”
  • “X software for the chemical industry”
  • “X software for the food industry”

Each industry gets its own subpage, so each page can be optimised specifically for it. That improves the chances of becoming part of the query fan-out, following this pattern:

  1. A customer asks, for example, ChatGPT for a solution and software for their construction company.
  2. Using query fan-out, ChatGPT runs background searches targeted at construction, for example:
    • “X software construction”
    • “X software construction finland”
    • “best X software for construction”
  3. The company’s landing page is optimised for exactly these kinds of keywords, so it ranks well for every one of these searches. Hyper-precise keyword targeting isn’t needed – that is already a bit too old-school these days.
  4. AI finds the content, reads it and identifies the brand, the product, its features and its target group.
  5. ChatGPT synthesises an answer from the content it found. Because this company is present in every search, it holds a strong position in the consensus from which the final answer to the user is generated.
  6. ChatGPT decides to recommend this company.

Worth noting: what the page actually says and shows is of course far from irrelevant. The content itself matters a great deal and can decide whether AI mentions and recommends you or not. That topic deserves its own article, though, so we won’t go deeper into it here. What matters most is high-quality content that speaks to and serves your target audience and works technically and structurally. As if you’d do it any other way.

AI optimization (in Finnish), then, is about understanding your positioning and the different contexts, and creating content online to match – both on your own website and on other sites. It’s worth understanding that a single site has only a limited ability to take over search results and exert influence.

This is also a good moment to remember that web pages are only one content format that can rank. Search engines increasingly show social media posts, videos and more, all of which can influence AI answers too. They affect some language models more than others.

Consensus picks the winner

When AI synthesises its final answer, it gives the consensus answer. It has no ready-made opinion of its own and knows no objective truth. AI uses whatever material it finds on the topic and has available.

A simple, simplified bit of maths helps illustrate this.

Assume a prompt triggers three background searches on search engines, and each search returns ten results. I say “assume” because we don’t know exactly how many results each language model goes through.

Amid all these assumptions, ten results per search is still a fairly reasonable figure: it is, for example, the number of organic results on a Google results page.

3 × 10 = 30 search results

So AI goes through 30 search results. A company should aim to own as many of those as possible. Three out of 30 may be enough for a recommendation, but the odds improve with more hits, say 5 or 10 out of 30. Especially if those hits talk about your brand with positive sentiment.

The three searches may all target the same context, such as construction, or they may spread across different searches. Then some of the results look very different.

In that case, winning the consensus clearly takes more content and more pages. A successful, durable strategy for AI visibility therefore requires a lot of analysis, digging and, in many cases, far more content than people imagine.

Keep it simple and get started

AI, its technology and how it works are highly dynamic. Updates come and go, new models arrive, and the weighting and preference for different sources change. Query fan-outs vary. New features come and go.

The most important thing is to understand the fundamentals and keep the building blocks of your strategy simple. Assume AI understands nothing unless you tell it yourself – exactly the way you want your story told. Take ownership of your narrative.

Position your company clearly and precisely. Create content that matches it, ideally across many channels, and you’ll start to command the consensus AI relies on. Make sure your content shows up in all the traditional search engines. That’s when you’re onto a winning strategy.

Jussi Viljanen

SEO consultant and entrepreneur who helps companies in Finland and internationally do more and better business through search engines and AI visibility. Over 10 years of experience with Finnish SMEs and in some of the world's most competitive industries.

More from this writer →

Read next