How one developer used Claude Code to build a geo-platform for brands across nine AI networks

One question in a work chat in September 2025 launched a full career shift for a mobile developer: he left his job at Biblio Globus to create a geo-platform that helps businesses monitor how their brand appears in responses from nine AI models. Two things matter in this story: demand for a new category of generative engine optimization and how much AI-tools like Claude Code have shortened the path from idea to product. According to the author, the trigger was a CEO's remark: "Who knows anything about GEO?"

After that, interest in the topic became concrete. GEO, or Generative Engine Optimization, is an attempt to adapt SEO logic to the era of chat assistants, when a user increasingly gets not a list of links, but a ready answer. For brands, this changes the task: it's not just about ranking in search results, but about appearing in the formulations, recommendations, comparisons and collections that models generate in dialogue.

At first, the author approached the idea as an engineer, not a theorist: he conducted a local technical audit of the frontend, observed how scenarios for interaction with AI models work, and in the evenings began building his own product. By that time, he had long followed the development of generative models, used ChatGPT from its first prominent launches, and from March 2025, according to him, became a daily user of Claude. So betting on Claude Code felt natural: the tool allowed for faster hypothesis testing, interface assembly, and moving forward without a team of several developers.

By December 2025, the side project had become his main occupation. The author left his job by mutual agreement and transitioned to full-time platform development. Its mission is formulated quite directly: to give business a tool that will take on the work of brand visibility in AI networks.

From this need arose the term GEO itself as a separate applied discipline. If companies previously measured traffic, positions and CTR, now they have to look at a new layer of metrics: does the model mention the brand in response to a user query, in what context does it do so, who does it show nearby, and what sources does it use when formulating the response. In fact, this is about transitioning from traditional web analytics to response analytics, where what matters is not just clicks, but meaning, formulations and the brand's place within AI dialogue.

From the article title, it follows that the platform works with nine AI models at once. This is an important nuance because the market stopped being a story about one leader long ago. Different models answer the same questions differently, rely on different web sources, and can give businesses completely unequal visibility.

For companies, this means that one optimization for AI is not enough: you need to compare responses between systems, track divergences, and understand where your brand has already established itself in recommendations and where competitors have taken its place. If one model confidently recommends a competitor while another doesn't mention your brand at all, this is no longer an abstract metric but a direct signal for marketing and content. In this context, the product solves not an academic but a very practical task of reputation monitoring, presence, and digital discoverability.

The main takeaway from this story is that GEO is gradually becoming a separate market, and the barrier to entry for creating B2B tools has noticeably lowered. One developer armed with a clear problem and a strong AI assistant can not just make a prototype, but build a specialized platform for a new user behavior model. For business, this is a signal that the fight for attention is shifting from classic search to AI interfaces.

For developers, it's a reminder that Claude Code and similar tools are becoming not a decoration to the process, but a real lever for launching products solo.