By Luke Marthinusen, CEO and Founder, MO Agency
The last 12 months have produced the biggest change in how customers find businesses since Google launched. Not a refinement of search. A replacement of it. And almost no one is talking about it with the urgency it deserves.
We've spent the past year optimising websites for AI visibility across our client base of enterprise and growth-led companies. What follows is what we've actually seen, what's working, what's not, and where the real difficulty sits, because it's not where most people think.
The new game is top three answers, not one of ten
For 25 years, the goal was to rank on page one of Google. Ten blue links, and if you were in the top three, you got the lion's share of clicks. That model is being replaced, fast, by something far more clinical.
When someone asks ChatGPT, Claude, Perplexity, or Gemini for a recommendation, the AI doesn't return ten options. It returns two or three. Sometimes one. The human reads them, picks, and moves on. They often don't even visit the websites anymore. The funnel that used to take days of browsing now collapses into a single conversation.
This is the part business owners need to internalise. There is no page two in AI search. There is no scrolling to find you. You are either in the recommendation set or not.
What we're actually seeing on client sites
Massive drops in organic traffic and a rise in direct traffic. This change reflects the drop in click-throughs from search engines. In most instances, click-throughs can't be tracked from LLM apps like Claude and ChatGPTs. That's why they pop up in direct traffic. If you've only seen a drop in organic traffic and no corresponding rise in direct traffic, you have an AI visibility problem.
Different AI systems trust different sources. This is the single most important insight we've developed, and it changes how you build a strategy. In our testing, Anthropic's Claude leans heavily on owned-domain content and press releases distributed through industry publications in your country or region. OpenAI's ChatGPT leans heavily on B2B directories relevant to your industry. Gemini behaves differently again. If you build a visibility strategy that feeds only one source type, you'll show up on one platform and be invisible on the others.
Client reviews unlock multiple platforms at once. Verified reviews on the right B2B directories for your industry are gold to the LLMs. Google reviews have also performed surprisingly well as a citation source in our testing. Reviews are among the highest-leverage actions a B2B company can take right now, and almost none of our clients were doing them consistently before we audited them.
The hardest problem nobody talks about: The prompt language gap
The visibility platforms (Profound, Getmd, Otterly, RankScale, and others now appearing weekly) all ask you the same thing: Which prompts do you want to track?
If you sell CRM implementations, you write: "Best CRM implementation partners on the East Coast."
The problem is that your customers don't use that language. They never have. They say things like "I need a CRM that doesn't cost a fortune for a 40-person sales team" or "How do I move off Salesforce." Those are the prompts that drive actual buying decisions, and they look nothing like the marketing-flavoured queries we instinctively reach for.
The way we now build prompt sets for clients is different. We pull from sales call recordings. We mine their support tickets. We look at Reddit threads in their category. We use the actual phrases buyers use when they don't know what to call the thing they want.
What to do in the next 90 days to improve your AI Visibility
If you want to be in the recommendation set when AI agents start doing more of the buying, and they already are, here's the priority order we'd run with any client:
One. Audit how often your domain is currently cited across the major LLMs for the queries that actually matter to your business. Not the queries you wish mattered. The queries your customers actually ask.
Two. Add structured data to every commercial page. Product or service schema, FAQ schema, Organization schema, review schema where you have them. This is a one-week project for most sites and the highest-leverage technical change you can make.
Three. Make sure the core information on your service and product pages is easy to extract. There's a balance to strike here between human readability and visual impressiveness on one side, and machine readability on the other. The information they need (what you do, who you do it for, what it costs, how it works) must be available in a clean, machine readable format.
Four. Build your off-domain footprint deliberately, with platform preferences in mind. B2B directory listings relevant to your industry (G2, Capterra, Software Advice, or vertical-specific directories, depending on your category). Press releases through outlets your buyers actually read. Client reviews on those same directories and on Google. The internet has to vouch for you before AI will recommend you.
The window is shorter than you think
AI search adoption is moving at a lightning pace. Gartner is forecasting that AI agents will be involved in a significant share of B2B and B2C buying decisions by 2028.
If you're running a business and you haven't audited your AI visibility yet, that's the place to start. Not because the change is coming, but because it's already happening, quietly, in conversations between your potential customers and the AI assistants you can't see.
The first three answers are the only ones that matter now. Make sure you're one of them.
Luke Marthinusen is the founder and CEO ofMO Agency, serving enterprise and growth-led companies internationally. MO Agency runs AI visibility programmes for clients across multiple industries.
COMTEX_482313994/2956/2026-05-28T12:06:54