--- title: "Manifesto | Obi" description: "Our thesis: AI search is the best go-to-market channel of 2027. Why our approach works, and our vision of Agent-Led Growth." canonical: "https://hireobi.com/manifesto" human_url: "https://hireobi.com/manifesto" markdown_url: "https://hireobi.com/manifesto.md" updated: "2026-11-10" --- # The antithesis of AEO tracking dashboards. AI search is the best go-to-market channel of 2027. This is our thesis, and here is why we hold it. ## I. Buyers now shop inside AI ### The shortlist moved. For twenty years, a B2B buyer searched Google, opened ten tabs, read reviews and booked a call. Today the first step is a conversation with ChatGPT, Gemini or Claude. The buyer describes the team, the stack, the budget and the problem. The assistant answers with three or four names. In a Semrush survey of US B2B professionals who use AI at work, 92% say AI has shaped their vendor shortlist and 83% say it influenced their final decision. 97% say AI helped them discover new vendors. Consideration and the shortlist, the two stages where a vendor is won or lost, now run through AI for every kind of B2B buyer. ### An LLM reads the buyer. A search engine ranks pages for a query. Two people who type the same words see nearly the same list, and that list rewards links, authority and years of content. An LLM reads the whole conversation: the role, the company size, the tools in place, what the buyer said three messages earlier. It splits the question into many searches, which Google calls query fan-out, reads passages from many pages, and writes one answer for that buyer. Ask twice and the answer can change. Ask as a different buyer and it changes again. So the brand that wins is the one that fits the buyer's situation. In the same survey, only 7% notice a vendor in an AI answer because they know its name; 53% notice it because it matches their use case. ### A young company can win fast. The common view says there is no GEO without SEO, and that known brands keep the advantage in AI answers. The buyers' own answers say otherwise. When AI finds no page that answers a buyer's precise situation, the first company to publish one gets a real chance to be named. With our clients and in our own experiments, we have seen it happen in weeks, with no SEO behind it. That is what makes AI search a go-to-market channel: it brings you buyers who look like your customers, at the moment they choose. ### This work is better done by an AI. AI search work is mostly reading: hundreds of buyer questions, on several engines, as several buyer profiles, every week, with answers that change between runs. Then your site, your competitors' pages, the reviews, the Reddit threads and your sales calls. A human team can sample a few questions a month. An agent reads all of it every week, the same way each time, and it reads the way a model reads. The human stays where judgment matters: you decide what goes live. Sources: - [Semrush, “How AI tools shape the B2B buying process”, July 8, 2026: 519 US B2B professionals who use AI at work, surveyed March to April 2026](https://www.semrush.com/blog/how-ai-shapes-b2b-buying/) - [Google Search Central, “AI features and your website”: query fan-out in AI Overviews and AI Mode](https://developers.google.com/search/docs/appearance/ai-features) ## II. Why our approach works ### It asks the questions buyers really ask. An LLM answers the conversation in front of it. A buyer brings a role, a team, a stack and constraints, and the answer depends on every one of them. Obi starts from your ICP and writes each question in your buyers' words, with their context, from sales calls, G2 and Reddit. It measures the answer your buyer actually reads. A generic question sent to a broad AI measures an answer nobody reads. That is spray and pray: a vanity metric that moves no pipeline and brings no revenue. ### It gives AI what AI is missing. To build an answer, an LLM searches, reads passages from many pages and keeps the ones that fit the buyer's situation. In the Semrush survey, 53% of buyers notice a vendor because it matches their use case, and 33% say AI recommendations are too generic for them. A page that answers a precise situation with facts only you hold fills a real gap. Obi finds that gap and writes the missing piece from your own data, in your voice. ### It is fast because the method is right. The sector says AI search takes time and effort, like SEO. Slow results come from the method: the wrong questions, asked for months. Work aimed at the exact buyer, the exact question and the exact missing fact gets picked up fast. That is why our clients see results in weeks. ### It counts revenue. A mention is a step on the way. Every question is scored by buyer impact, and Obi reports what moves: users, demos, revenue. This is the work of an employee, so we built one: Obi, working right from Claude Code, Codex, Cursor or any terminal. Sources: - [Semrush, “How AI tools shape the B2B buying process”, July 8, 2026: 519 US B2B professionals who use AI at work, surveyed March to April 2026](https://www.semrush.com/blog/how-ai-shapes-b2b-buying/) ## III. Our vision: Agent-Led Growth Companies grow through sales or through their product. We believe every company must now also grow through agents. Agents already choose for people. ChatGPT and Gemini build the shortlist, coding agents like Claude Code and Codex pick the tools a project uses, and personal assistants like Dots and Muse act for their users. That holds for consumer, B2B and enterprise companies, whether their market is narrow or immense. If LLMs, coding agents and personal assistants are not on your side, you will lose the battle, even if you hold a monopoly today. We call it Agent-Led Growth, and Obi is where it starts. This is our thesis and our vision. Nazim Zidi, Founder (https://linkedin.com/in/nazim-zidi/), 2026-11-10 ## Next actions - Hire Obi, your AI Search Employee: https://agent.heysurface.com - How Obi works: https://hireobi.com/method.md