There is a brand you'd never heard of three months ago that today AI systematically cites in the very same queries where you want to appear — and it shows up in none of the SEO reports you're looking at. By the time you notice, it's often already too late: it has already claimed the space and built the signals that matter. Monitoring new entrants in AI answers lets you intercept them while they're still small, before they become the reference point of the industry.
For three months a brand has been showing up that AI cites more and more often in your industry. You don’t know it, it wasn’t in your SEO reports, you can’t even find it on the second page of Google. Yet ChatGPT names it, Perplexity puts it in its sources, Gemini suggests it when the user asks for advice.
It’s a new competitor, a shift in strategy by a small brand, a company that has understood something about the way AI engines select their sources. You don’t know which. And you never will if you keep looking only at Search Console and SEMrush.
Let me explain how to spot it in time, why it matters for your visibility in AI answers, and what to do once you’ve identified it.
What I mean by new entrant in AI answers
A new entrant, in AI monitoring, is a brand that until 60-90 days ago never appeared in the answers of AI engines for your industry queries, and now appears with increasing frequency. It isn’t necessarily a new company: often it’s a brand that already existed but that has crossed a visibility threshold inside the generative models.
The point is that AI engines don’t behave like Google. Google shows you the sites that rank for that keyword: a small brand without backlinks struggles to reach the first page. ChatGPT and Perplexity, on the other hand, draw from very different sources — specialist articles, vertical forums, citations in journalistic pieces, listings on niche directories — and they can name a brand that on Google sits on page 5.
In previous articles I talked about how implicit citation mechanisms work and about why backlinks count as a citation proxy. Those mechanisms explain why a small brand, well cited in thematic contexts, can appear in AI answers before it appears in the SERPs. From this follows a consequence worth stating clearly: monitoring only the SERPs leaves you blind to a substantial part of the competition that is taking shape.
Why AI monitoring sees things classic SEO tools don’t
Traditional SEO tools think in terms of keywords and rankings. They’re perfectly fine for their purpose, but the conceptual model is “who is in the top 10 for this query?”. AI answers, instead, compose different sources, cite brands within lists, recommend products without a single clear starting query.
The practical result is that a competitor can gain AI citation share for months without moving an inch in SEO rankings. When you finally see it growing on Google, it has already consolidated a presence in hundreds of AI answers generated every day. The competitive advantage has already shifted.
This is why, in the monthly monitoring of AI visibility, it’s worth tracking every brand cited, not just the ones on your starting competitor list. The competitor list is the one you built yourself, based on your perception of the market. The list of brands AI cites is the reality the model is putting in front of users.
A single measurement is a snapshot, it tells you nothing about trends.
The test you can run in 30 minutes
Let me tell you how I do it when I want to understand whether a new brand is emerging in a client’s industry within AI answers.
Take 8-10 informational queries that a potential client actually asks. Imagine a small producer of alpine liqueurs in Piedmont who makes mountain-herb amari: the queries of one of his potential customers might be:
- “best artisanal Italian amari with alpine herbs”
- “traditional liqueur producers Italian Alps”
- “what to give a lover of mountain amari”
- “differences between genepì and alpine amaro”
- “artisanal distilleries in Piedmont to visit”
You run them on ChatGPT, Perplexity, Gemini and Claude. For each answer you note down every brand cited, even in passing, in a table. Brand column, engine column, date column. Repeat the operation once a month, always with the same queries, always with a “clean” account (incognito mode or a dedicated profile to avoid personalization).
After three months you have a matrix. The brands you see growing month over month are your new entrants to investigate. The ones that disappear are the brands losing ground in the models’ internal ranking.
For industry queries you can lean on Google Trends to understand whether the term is gaining volume, and on your site’s Google Search Console to see whether impressions on that query are growing or declining in parallel. They don’t solve the AI detection problem, but they give you the SEO background context.
You need at least three observation cycles, spaced 30 days apart, to distinguish a random fluctuation from a real growth pattern.
The test I ran myself
For three months I followed a cloud of queries in the artisanal alpine liqueurs sector, for a Piedmontese producer who sells amari, genepì and ratafià to a small network of wine shops in the North West. Stated goal: understand which brands AI cited in that sector and whether the list changed over time.
Set: 12 informational queries, 4 AI engines (ChatGPT, Perplexity, Gemini, Claude), one measurement every four weeks, always the same day and the same time slot. Three cycles in total.
Pattern observed: 14 brands cited at least once in cycle 1. By cycle 3 there were 19. Three brands from cycle 1 had disappeared. Eight were stable. Eight were new entries. Of those eight, two appeared with increasing frequency across three different engines (not sporadic): a small Piedmontese producer who had opened a very well-curated thematic blog, and a mountain microbrewery that had launched a line of liqueurs and earned two reviews on regional food-and-wine publications.
Limitation of the test: the query sample isn’t large, with no claim of statistical representativeness. It’s an indicative test that shows the method, not a study. Real analysis for a brand with a budget requires professional tools with continuous tracking, automatic alerts, comparison across multiple engines and clustering of similar prompts.
But the qualitative signal was clear: in 90 days the landscape of brands cited by AI in the sector had changed by about 30%. Had that producer kept looking only at his three historical competitors (the same ones he monitored on SEMrush) he would have completely missed the arrival of the new ones.
The most common mistakes in competitive AI monitoring
Tracking only known competitors. You build your Excel with the 5 brands you consider competitors, you check every month whether AI cites them, you see the situation is stable and you feel reassured. But you’re staring at the finger while ignoring the moon… Yes, because what matters is the movement of the brands outside the list.
Taking the measurement only once. A single measurement is a snapshot, it tells you nothing about trends. You need at least three observation cycles, spaced 30 days apart, to distinguish a random fluctuation from a real growth pattern.
Mixing personalized accounts and neutral sessions. If you query ChatGPT from your account that has been talking about your industry for months, the model knows what interests you and might give you richer answers than usual. For competitive monitoring you need the answer a generic user receives, so incognito sessions or dedicated profiles.
Considering only the brand cited by explicit name. Sometimes AI cites “an artisanal Piedmontese distillery that produces high-altitude herb amari”: it isn’t a brand by name, but it’s a semantic citation that often resolves into a specific brand if the user asks to dig deeper. It’s worth noting down these descriptive patterns too.
What to do once you’ve identified a new entrant
When, after three months, you see a brand growing steadily in AI citations in your industry, there are three things to do, in order.
- Profile it from the outside. Visit the site, read the latest 5-10 pieces of content, check the “about us” page, go to Wikidata and search whether it has an entry. Understand who it is, how long it’s been around, what it talks about.
- Find the sources that cite it. Ask the AI itself: “which sources do you cite when you talk about [brand]?”. It often gives you 3-5 direct links — articles, reviews, directory listings. Those are the sources that generated its AI visibility. They’re the same ones you should target for your own brand.
- Compare its semantic footprint with yours. Which entities does it name in its content? Which people does it cite? Which specific events/places/products? In previous articles I talked about named entity recognition and about event entity speaking authority: the density and specificity of the entities named is one of the strongest signals for why a brand enters AI answers before others do.
Always compare with the 3-5 competitors AI cites consistently in your industry, not just with the new entrant. The combination of where you stand relative to the stable ones and what the newcomer does differently gives you the map to recover ground.
Why new entrant detection is an integral part of visibility in AI answers
Visibility in AI answers is not a fixed position you win once and then keep. It’s a dynamic balance between you and the set of brands the model considers relevant for the queries in your industry. Every new brand that enters marginally shifts the space available for the others.
Noticing in time who is coming in lets you understand which sources, which content, which semantic patterns are working right now for that sector. It’s information worth having to redesign your presence, not just to play catch-up.
In the next articles in the series I’ll tell you how to structure a monthly matrix of AI share of citation between you and the competitors, how to automate part of the measurement with basic scripts, and how to cross-reference AI citation data with traditional brand awareness indicators to understand whether a new entrant’s growth is sustainable or just a flash in the pan.
