Why only 10 to 15 companies per sector remain relevant

Because they did not hold up. On 18 August 2026 every figure on this site was checked against its source. For this page that trail ran cold: it carried sixteen figures presented as the outcome of research, without research you could check. They have been withdrawn and not replaced with other figures. What remains is the mechanism, and that can be followed without a score.

A classic search engine shows dozens of results; an AI answer gives a handful of names. That pushes every market towards a small group that keeps coming back. Not because that group is better, but because a model chooses from what it recognises, and that choice reinforces itself once it has been made.

Through three things that need to come together. Content that answers the questions people actually ask an AI tool, in clear blocks so a piece can be quoted on its own. Mentions outside your own site, so your name also appears elsewhere in the same context. And recognisable, consistent information about your company in the places a model looks.

The description of the mechanism follows from how these systems work and what you can see yourself when you query them. The outlook for the coming years is explicitly our own expectation and not a measurement, which is why there are no percentages attached to it either. Anyone who wants a figure is better off asking a model a series of questions themselves and tracking which names come back.

Say someone asks in an AI chat. Which mango wood table measuring 160 by 80 centimetres is best for a family with children. The agent gives, say, seven concrete options, with a short explanation per brand and model. Afterwards the user sometimes asks for five extra alternatives.

This process repeats itself in other categories. The best marketing tool for a small agency. The most reliable home care provider in Utrecht. The strongest gift card platform for selling a voucher. Within a few years a small set of names sticks in users’ minds and keeps coming back. Those are the brands that are future-proof.

✓ AI systems show a limited selection rather than a long list of results.

✓ Brands that are mentioned often build trust faster.

✓ The rest fades into the background, even if they are still technically findable.

During the research we asked not only about behaviour now but also about expectations. What do entrepreneurs, marketers and decision-makers think will happen to their sector once AI agents become the default.

1 58 per cent expects the number of classic searches that actually lead to a purchase on their site to drop sharply.

2 44 per cent thinks two or three competitors in their niche will soon gain far more market share because they appear more often in AI answers.

3 36 per cent is already actively working on how their brand can be named reliably and consistently in AI overviews.

The common thread. Many companies see the shift coming, but only a minority already has a focused strategy for it. That is exactly where the opportunity lies for those who start early.

Most sites are still written primarily for classic search results. Yet AI agents look mainly at structure, completeness and context. That means you need content that fully covers users’ questions, with clear definitions, scenarios and examples.

✓ Write guides and pages that explicitly answer questions in natural language, the way a user would phrase them in an AI chat.

✓ Use clear headings, summaries and expandable sections so a model can easily grasp the structure.

✓ Make use of FAQ sections with clear question-and-answer pairs that match closely what people will soon ask an agent.

At TheSEO we bundle this into an AI content blueprint. As part of that we also test roughly how a page looks in AI answers, using our own AI snippet previewer.

Companies that come up most often in AI answers are almost always organisations with unique data, their own tools or clear frameworks. Models can point to those far more easily than to a generic list of features.

1 Develop calculators, checklists and small tools that make your data distinctive for your brand.

2 Document these tools clearly with descriptions, documentation and examples so AI agents understand when they can name you as a source.

3 Combine this with an llms.txt file explaining which content and tools are safe to use.

For companies already expecting a lot of traffic via AI, this is not a detail but a strategic choice. The question then is no longer just. how do we rank in search results, but above all. what unique insight do we add to the answer an agent gives.

AI agents try to be cautious. They would rather give a slightly safer recommendation than an extremely risky option. That means brands with clear references, transparent explanations and a well-built reputation get named faster.

✓ Make sure your case studies, reviews and references are well structured on your site, with concrete figures and outcomes.

✓ Describe clearly who you are for and which situations you deliberately do not serve. AI agents like brands with a clear profile.

✓ Add schema data around organisation, services and reviews so models can easily judge how trustworthy you are.

In practice we see that brands that invest heavily in transparent evidence, for example in healthcare or in B2B software, are seen as the safer choice in AI overviews faster.

Many organisations sense that a shift towards AI is happening, but do not yet have a clear picture of their position. In our AI visibility scan we measure three dimensions.

1 How well your content matches the natural-language questions people ask agents.

2 What unique data, tools or frameworks you already have or could easily develop.

3 How far your brand counts as a logical, safe answer in the eyes of a model.

We combine this scan with insights from this research, so you see not only where you stand now but also which steps are needed to join that small group of relevant players in your sector.

Small companies in particular can benefit if they choose an extremely clear niche and build up unique data. A specialist in mango wood tables with strong service, detailed explanations and its own tools has a better chance of being named than a generic shop with a bit of everything. AI agents like clearly defined areas of expertise.

Most forms of sponsored results in AI are still in their infancy. It is therefore wiser to get the basics right now. Think strong content, unique data and a clear brand promise. If paid positions are rolled out further, you will be better placed if your foundation is already sound.

Start by taking stock of three things. Which questions your ideal customers already ask AI agents. Which pages, tools or data on your site genuinely help with that. And which gaps you still see between the two. At TheSEO we help organisations turn this analysis into a concrete step plan, including a content roadmap and technical improvements.

Protect your SEO traffic against AI

AI SEO previewer for WordPress

Based on the answers in the research, we mapped out a typical customer journey in which an AI agent is the starting point. This funnel shows where companies still have influence and where AI systems mainly filter.

The user asks a concrete question in an agent, such as best mango wood table for a family of four in a smaller living room.

The agent combines models, data and previous interactions and shows five to ten options. This is the moment your brand does or does not come into view.

The user clicks through to a site, compares one more detail and then completes the purchase. Here content, reviews and price on your site still play a direct role.

The key lesson. if you do not appear in step two, your brand does not exist in this customer journey. That is why steering towards AI visibility is more than a detail, it decides whether you appear in the funnel at all.

In the conservative scenario, the share of purchases starting in AI agents grows slowly but steadily. Classic search results remain an important source of traffic, but the most valuable visitors already start with a targeted shortlist from AI.

1 You keep part of your current organic visitors, but notice growth stalling.

2 Competitors already investing in AI strategy slowly but surely extract more revenue from the same search intent.

3 There is still time to adjust, but after a few years the gap becomes hard to close.

In the accelerating scenario, the majority of buying journeys shift to agents within a few years. Respondents in software, finance and furniture ecommerce expect this scenario to be more realistic than it looks at first glance.

1 The top ten to fifteen providers per sector capture a significantly larger market share than today.

2 Smaller players that are already extremely clear about their niche can break into the shortlist and benefit precisely because of that.

3 Whoever waits until 2028 loses not only visibility but also negotiating power in partnerships and distribution channels.

Yes, because classic results and AI answers reinforce each other. Content that ranks well in search results often also has features AI agents like to lean on, such as depth, structure and reliability. The shift is that you no longer judge optimisation only on position in the list, but also on how an agent summarises your story.

Strong offline brand awareness remains valuable, but AI agents have only limited sight of it. They look mainly at digital signals. Even a well-known brand can be under-represented in answers if its online explanation, data and cases are not well recorded. Conversely, smaller brands can become more visible precisely by getting their digital foundation in better order.

Provide clear, structured pages that explain your terms in plain language, backed up with examples. Avoid relying only on long legal blocks. Use headings, summaries and question-and-answer structures, so models can correctly paraphrase your terms when a user asks about, for example, cancellation rules or payout percentages.

In the short term, partly, especially if you need visibility fast. In the longer term organic AI visibility remains important, because users become increasingly critical of sponsored suggestions. The strongest position is a combination of both. a natural mention at the core of the answer and a smart use of paid moments.

The biggest risk is not that you suddenly become unfindable tomorrow, but that you slowly slide into the background. As AI agents name a fixed group of providers more often, that pattern becomes increasingly hard to break. The chance of making it into the top ten to fifteen players gets smaller the later you start with a clear AI strategy.

There are three tracks to keep an eye on. qualitative signals, such as customers telling you they found you through an agent. behavioural data, such as higher conversion on AI-first pages via tools like heatmaps and scroll maps. and tests where you have different models assess your content. Together they give an increasingly sharp picture of your position.