Source: Unite.AI
Gintare Rimolaityte is the CEO and co-founder of Trendos, the AI search visibility platform that helps brands understand and optimize how they appear across LLMs. She gained early experience at Oxylabs in 2020, first as Business Development Manager and then as Head of Commerce, where she led Sales, Business Development, Account Management, and Customer Success teams for one of its business units. From 2022 until 2025 she served as CEO of engagement platform JumpTask. Prior to co-founding Trendos in 2026, she worked briefly as a Product Manager at Adment, focused on affiliate-fraud prevention for paid search and affiliate programs. As Trendos’ CEO, Gintare manages multiple teams and oversees strategic growth. She lives and works in Vilnius.
Trendos is an AI search visibility platform that helps brands monitor, measure, and optimize how they appear across LLMs and AI search engines. Using Trendos, brands eliminate the major blind spot created by the shift from Google search to conversational AI, which made the recommendation process practically invisible to them. Beyond tracking overall visibility, Trendos makes AI market intelligence tools accessible, intuitive, and affordable so brands can influence and optimize when and how they appear. Its technology captures the real answers and ads shown within AI environments through large-scale scraping of LLM sessions, then analyzes and delivers structured data as clear insights both into brand visibility and competitor activity. Trendos was founded in March 2026 in Vilnius, Lithuania and backed by Tesonet, a global venture builder behind companies such as Nord Security, Surfshark, and Hostinger.
Your career has taken you through commercial leadership at Oxylabs, running JumpTask, and product work at adment.ai before co-founding Trendos. What changes were you seeing in how consumers discover brands through AI that convinced you there was a need for Trendos, and what did you believe existing SEO and AI visibility platforms were missing?
Back when I regularly conducted interviews with marketing managers, I heard the same thing over and over: AI search is coming, and although there are a few tools out there already, nobody really knows how reliable they actually are.
I didn’t just have to take their word for it, since I was also watching the shift happen in real time. Consumers were moving from scrolling through links on Google to just asking LLMs directly. The latter was a bit of a black box for marketers. They couldn’t see if they were even being mentioned, let alone how their brand was being described or positioned relative to competitors.
I saw two problems with the existing software. First, many of the tools felt like SEO platforms with a bolted-on AI feature. Second, the few purpose-built tools that did exist were priced for enterprise only. If you weren’t a huge brand with the budget to figure out its AI visibility, you were locked out and flying blind. This didn’t sit right with me. Every brand deserves to know how it’s showing up, not just the ones able to shell out five figures a month. With Trendos, we’re intent on giving any size brand an accurate and fairly priced way to understand and optimise their AI visibility.
AI-generated answers can appear like a black box to marketers. From the data Trendos is collecting, what are the most important signals that seem to determine whether a brand is mentioned, cited, or recommended by systems such as ChatGPT, Gemini, Claude, and Perplexity?
Signal-wise, it’s really about being mentioned enough in the right places and in a structured way. That’s different from SEO, where you need a do-follow link for authority to transfer. With AI search, models aren’t indexing your site, they’re scraping it. The link becomes superfluous. You just need your brand name to show up clearly and often on the pages these models already pull from.
From our data, five pillars feed into this: your own website, PR and earned media, review platforms like Trustpilot and G2, affiliates, and social. How much each one matters varies by industry. Apparel and e-commerce lean on owned media, since product pages are easy for models to extract. Consumer finance leans more on affiliate comparison sites. SaaS and B2B brands get cited heavily from review platforms like G2 and Capterra. There’s no universal formula. You have to look at your own citations to see which pillar matters most for you.
Structure matters too. The brands we see cited most often publish clear, direct content that answers questions up front, often in an FAQ style. That’s easy for models to extract. Keyword-stuffed blog posts don’t help. What works is answering the question first and adding detail after. On AI-generated content, we don’t currently see models penalising it. What gets cited is content that answers the question well, regardless of who or what wrote it.
There’s some randomness to the whole endeavour too. These models predict the next word rather than run one fixed ranking algorithm. You can do everything right and still see volatility. At Trendos, we try to help brands focus on the signals they can control, and quit chasing the ones they can’t.
Unlike a traditional Google ranking, an AI answer can change when the same question is asked multiple times. How do you measure AI visibility reliably when the underlying models are probabilistic, and how much sampling is necessary before a brand can draw meaningful conclusions from the data?
This is a hot topic in the industry right now. One company recently took to LinkedIn to argue you need to scrape an answer around 100 times a day and average the results to get something reliable. Some of the bigger platforms still feel once a day is enough. I don’t think either side has fully backed up their position with real evidence.
We look at it empirically. We have clients like NordVPN who’ve been monitoring with us for some time now, and we can see the volatility in their visibility and rankings over that extended period. It just doesn’t move around that much. Scraping more frequently probably buys you a bit more precision, but the margin is a fraction of a percent and certainly not the difference between being invisible and being number one. It might shift someone from 35th to 32nd place, which matters much less than the bigger structural things a brand can actually act on.
Our approach is to monitor consistently over time rather than chase extreme sampling frequency, and let brands look at their own trend line to judge how stable or volatile their category is.
There is considerable debate around SEO, answer engine optimization, and generative engine optimization. Which principles from traditional SEO still matter in AI search, and where do you think marketers make a mistake by simply applying their existing SEO playbook to large language models?
A number of SEO principles still matter in AI search. Take query fan-outs, for instance, which are the follow-up searches ChatGPT runs behind the scenes before it provides an answer. Those still come from real search behavior, making the associated ranking an SEO exercise. You’re still optimizing long-tail keywords, only aimed at what the model searches for rather than what a person types into Google.
Lots of marketers mistakenly assume a mention needs to work like a backlink. In SEO, you want a do-follow link because that’s how authority and domain rating actually transfer. In AI search, the models are not indexing your site. Since they’re scraping it, so you don’t need a link at all, just clear enough brand mentions on a page the model already trusts. People spend a lot of time chasing link placements when what they should be chasing is the mention itself.
Another common mistake is carrying over old SEO habits. Content written mainly to hit keywords, with the actual answer buried halfway down the page, tends to get overlooked by AI models. Answer the question first, clearly and directly, then add the detail.
I would say it’s also a mistake to decide SEO is dead and start from scratch. SEO isn’t dead! GEO is an extension of SEO, plus social and affiliates all working together. Brands that already do SEO well haven’t done all that work for nothing, but they also shouldn’t assume the old rules apply one-to-one.
Trendos tracks the sources that AI systems cite when answering questions about brands. What have you learned about the types of sources AI engines appear to trust most, and how much influence can a company realistically have over those citations without trying to game the system?
Trust isn’t a single hierarchy. It depends on the industry. What is cited most in apparel (owned media) isn’t what’s cited most in consumer finance (affiliate content and apps). No single source type wins everywhere. You have to look at your own category’s citations to see what’s being trusted.
Structure matters as much as the type of source. Authoritative editorial sites still carry weight, but it’s not about a single authority score like domain rating. What we see is that consistency matters more: how often a brand gets mentioned across a mix of trusted sources. And a clear, well-structured page from a smaller site can outperform a long article from a bigger publication, simply because it’s easier for the model to extract and use.
As for influence without gaming it, I’d separate this from an SEO mindset entirely. You’re not trying to manipulate a link graph. You’re trying to earn genuine mentions across those five pillars: your own site, PR placements, review platforms like Trustpilot and G2, affiliate relationships, and social presence. Then structure everything so it’s easy for a model to read. That’s not that different from good brand building. The most visible brands focus less on gaming the system and more on being verifiably present in the places models already trust.
Trendos launched an index tracking the 1,000 brands with the greatest AI visibility across different markets. What patterns or surprises have emerged from that dataset, and what separates brands that consistently surface in AI-generated recommendations from competitors that remain largely invisible?
The first surprise is how differently the engines behave. In the same market, ChatGPT and Google AI Overviews share in their recommendations only around two-thirds of the country’s top 100 brands. AI Overviews leans heavily towards everyday retail. Trader Joe’s ranks 35th there in the US, but 567th on ChatGPT. ChatGPT leans towards B2B and enterprise brands like Gartner, ServiceNow and Databricks, which AI Overviews largely ignores.
The second is how local it is. Seventy percent of the brands in our index appear in only one country, and in Germany almost a third of the top 100 are brands you won’t find in the UK or US lists at all.
What separates the consistently visible brands is breadth. A typical top-10 brand shows up across more than 90 product categories; a brand ranked in the bottom half shows up in about five. They’re not just mentioned more often, they’re relevant to far more questions. And the lists move fast: the very top barely changes month to month, but around 13% of brands drop out of each list within a single month. Visibility isn’t something you win once; you have to keep earning it.
Trendos’ Ad Radar tracks sponsored placements appearing alongside ChatGPT conversations. What are you seeing in the early data about which industries and advertisers are moving fastest, and how should marketers distinguish paid visibility from brands being organically mentioned or recommended by the underlying AI model?
ChatGPT ads skew heavily retail and B2C, making up roughly 50 to 70% of what we’re tracking. The rest is B2B, mostly companies and agencies advertising their own services. That makes sense, since ChatGPT’s user base is predominantly consumers, but with a real B2B presence too.
It’s still early days for OpenAI’s ad platform. It keeps adding new formats and features, and the rollout is uneven: when I last checked, ads weren’t live or billable in several European markets, even after OpenAI announced its expansion there. So any read on the fastest-moving industries is more a snapshot of who’s willing to test something new than a picture of who’ll still be advertising in six months.
That’s also where the gaps are for advertisers. OpenAI doesn’t show you which prompts your ads appear on, and it doesn’t show you what your competitors are doing. With Trendos’ Ad Radar, brands can see exactly which prompts trigger their ads, and see the other advertisers in their category: how many ads they’re running, where they land, and which headlines and descriptions they use.
For now, paid and organic mentions are easy to tell apart. ChatGPT clearly labels sponsored placements and shows them at the bottom of the answer, so nothing is blended into the response itself. An organic recommendation is the model choosing you as the answer; a paid placement is you renting a spot below it.
My honest expectation is that this will get harder over time. The more an ad looks and feels like a genuine part of the answer, the better it tends to perform. That’s been true on Google and Meta, and ChatGPT will likely move in the same direction. Which is exactly why brands need to track both their paid and organic visibility side by side.
One of the challenges with generative engine optimization is separating correlation from causation. When a brand improves its AI visibility, how can a marketing team determine which action actually caused the improvement, whether that was digital PR, third-party coverage, stronger first-party content, reviews, technical changes, or something else?
Attribution is one of the hardest questions in GEO, and nobody can yet draw a perfect line from every action to every result. But brands are far from guessing. With Trendos, you can see exactly which sources AI engines cite when they mention you and your competitors, and that tells you a lot about what’s working.
The most reliable approach is to connect a specific action to a specific source. If you land coverage on a site that’s already cited in your category, you can watch whether that source starts appearing in answers to the relevant prompts and whether your brand comes with it. The same goes for reviews, affiliate placements or new content on your own site. You’re not waiting for your overall score to move and wondering why; you’re seeing which sources the models actually use.
Paid is even clearer. With ChatGPT ads, you can track direct impact the same way you would any other paid channel, using UTMs to measure traffic and conversions.
It’s still worth looking at trends over several weeks rather than reacting to a single day, because AI answers naturally vary. We’re also building Key Insights, which surfaces these connections automatically, so teams don’t have to dig through every graph themselves. There’s still some uncertainty in GEO, but brands now have the tools to see clearly what’s giving them the most value.
Trendos recently introduced an MCP integration that allows AI agents to query visibility, competitor, citation, and advertising data directly. Do you see AI visibility management eventually shifting from marketers studying dashboards to agents continuously identifying gaps and recommending or executing changes on their behalf?
I do think that’s where this is heading, and we’re already building toward it. Even without agents in the picture, Trendos is moving toward doing things for brands rather than just showing them a dashboard. We recently found a small producer whose website was completely blocked to AI bots due to a basic technical issue nobody on their team knew to check. We told the brand about the problem, but our ambition is to be able to fix it for them without someone on their side needing to interpret a graph first.
Agents are a natural extension of that. If a system can already read the data behind Trendos, it shouldn’t be a huge leap for it to then flag a missing citation and act accordingly. I think dashboards won’t disappear entirely. Our more technical users still want the raw data, especially the exact query fan-outs and the actual answer text. But for most brands—the 90% who don’t want to be data analysts—I do think the endpoint is an agent quietly managing this in the background and only surfacing it when something needs a human decision.
As AI assistants increasingly move beyond answering questions toward recommending products, comparing vendors, and eventually completing transactions through agents, how do you expect the concept of “search visibility” to evolve, and which metrics do you think brands will ultimately care about most?
I think “search visibility” as a concept will become one part of a bigger picture, which is whether an agent chooses your brand when it’s acting on someone’s behalf. Being mentioned in an answer is one thing. Being the brand an agent actually selects, compares favourably, and completes a transaction with is a different, much higher bar. We’re already seeing the beginning of this with agentic commerce.
Once agentic commerce takes off, visibility alone won’t be enough. Sentiment will skyrocket in importance. An agent comparing three vendors on your behalf is going to weigh not just who’s mentioned, but who’s described as trustworthy, well-reviewed, good value. The recommendation rate—meaning how often you’re the one selected out of the brands considered—will be a far more meaningful number than raw mention count.
The other thing I’d expect to matter more is structure and machine-readability. Right now it’s about writing FAQ-style content so a model can extract it easily. Once agents are transacting on your behalf, it’s going to be about whether your pricing, your inventory, your terms are structured in a way an agent can actually read and act on directly, not just describe.
An important caveat: it’s early. I’d rather say that plainly than pretend we have this fully mapped out. What I am sure of is that the brands paying attention to this shift before agents are transacting at scale are going to have a real head start, much the same way the early SEO brands did with Google.
Thank you for the great interview, readers who wish to learn more should visit Trendos.