Georgia Leads Region in AI Adoption, but Companies Risk Being Invisible to AI Search

Why AI trusts media and independent sources more than company websites

Companies can remain invisible in generative AI responses even with good Google rankings. A growing body of research concludes that generative models cannot be 'bought' with the same tools that worked for classic search. Kristina Vorobieva, founder of the communications agency Sparrow PR, explains what influences AI search results and how Georgian companies can get into them.

Georgia is a telling example of how quickly this shift is happening. According to the Global AI Diffusion Report by Microsoft, in the first quarter of 2026, the share of Georgia's population using AI services (AI User Share) reached 17.26%. This is above the global average of 15% and close to neighboring countries—Azerbaijan (14.23%) and Kazakhstan (13.5%).

Businesses are adopting AI no less actively. A survey by the Georgian Business Association (BAG) showed that almost half of surveyed Georgian companies are currently in the process of implementing AI to boost productivity, and 13% have already fully integrated AI into one or more business processes.

Together, these data mean that the most active and commercially significant part of Georgia's audience has already begun shifting from traditional search to AI, and it is happening quickly. According to the 'Trends and Predictions 2026' report by the Reuters Institute, referral traffic from Google to publisher sites fell by a third over the year: users increasingly get answers directly from AI assistants without clicking through. This trend is likely to intensify.

For AI, what matters is not the size of a company's optimization budget, but the level of trust built by independent sources, says Kristina Vorobieva. In her view, this opens new opportunities for companies that cannot compete with major players in marketing budget size.

Why AI sees your brand differently than a search engine

While search engines rank pages, generative models form answers based on sources they have learned to trust, and that trust is built almost entirely outside the brand's own website. An analysis by Muck Rack of what ChatGPT, Claude, and Gemini actually cite showed that about 82% of citations are earned media (organic media publications), not the brand's own content.

Several other studies confirm the same pattern—including GEO platforms Zyppy, Authoritas, and Profound. Presence in authoritative publications—earned media—turned out to be the strongest predictor of citation in ChatGPT, with an effect of about 325%, according to AuthorityTech.

Thus, authoritative publications like Reuters and Forbes are among the top 20 sources cited by AI, alongside major platforms like Reddit and YouTube, as follows from an analysis by the GEO monitoring platform Similarweb.

In this way, generative search begins to evaluate the reputation of a source, not technical optimization. Traditional search rewards investments in more expensive advertising and tougher competition for positions in results. In generative search, a small company can appear in the same AI answer as a market leader if its experts consistently publish and are cited in authoritative media. Trust cannot be bought through paid placements; it can only be built, gradually and deliberately.

What AI trusts and what companies should do right now to get into the results of Claude, ChatGPT, and Gemini.

Preparing for AI search does not mean abandoning SEO. It means expanding the communication strategy beyond one's own channels. Before launching SEO or GEO promotion, it is important not only to align information on your own platforms but also to ensure basic brand visibility in external sources—media, industry blogs, podcasts, ratings, and professional communities.

To become visible to AI with a limited budget, companies should focus on several things:

Build a digital information trail through regular media publications. Recent studies show that content update frequency is a key factor in citation likelihood, and the exact list of sources AI cites for a specific query can change from one search to another. However, sources that AI models consider reliable and authoritative tend to repeat consistently across different results. By investing time and resources in publications in quality media, companies build trust with both clients and AI systems—trust that cannot be bought with an advertising budget.

Invest in developing the personal brand of leaders (thought leadership). In the era of generative search, an expert becomes as much a digital asset as the corporate brand itself. When an expert's name regularly appears alongside a specific topic and company name, a stable association forms between the person, brand, and area of expertise, so interviews with founders and authored columns work much longer than a press release. An analysis by the SEO platform Semrush showed that E-E-A-T signals (experience, expertise, authoritativeness) are the second strongest predictor of citation by AI systems, second only to the clarity and structure of the answer itself.

Maintain consistency of messages on the website, in interviews, and on social media. One of the most common mistakes is different positioning on different platforms: one message on the website, another in the media, a third on social media. To an algorithm, this looks like information noise. It is worth defining three to five key messages and using them consistently everywhere, from blog posts to media publications.

Build an information trail. Information independently confirmed in multiple sources inspires more trust in AI systems: if the same fact independently appears in media publications, research, and industry blogs, the model is more likely to consider it reliable, whereas a brand's claim about itself does not carry the same weight. The previously mentioned University of Toronto study confirms this: generative models systematically prefer independent publications to brand websites regardless of industry, query language, or question wording. For example, in the software category, Google referenced company websites in 43.7% of cases, while AI systems took 72.7% of sources from independent publications and only 26.7% from the brand's own website.

Publish content that directly answers customer questions. A separate section with answers to the most frequent customer questions helps a brand become part of the AI answer itself, not just a link next to it.

'AI does not choose whom to trust from scratch,' says Kristina Vorobieva. 'It accumulates trust that has already been built by media, experts, and every independent source that has already vouched for you.'

SEO taught brands to fight for a place in the results. GEO forces them to fight for a place inside the answer itself, and the winner is not the one with the biggest budget, but the one who has earned the trust of independent sources, one publication at a time.

Reference: Kristina Vorobieva is the founder of the communications agency Sparrow PR, an expert in communications with 18 years of international PR practice, including leadership positions at Havas, MINT (BBDO), and Saatchi & Saatchi.