.png)
Here is what our AI search data reveals and how brands can use it
The future of search is here and it’s not even text on a search bar. It’s a real-time log of something entirely new: prompted AI sessions. At Limy, these fresh approaches are a window to the past two decades of SEO and inspiration into a new metric – GSEO (Generative Engine Optimization).
What is GEO? And why does it matter?
GEO goes beyond keywords and reveals rich, useful content that shapes people’s outcomes and next steps. Compared to user planning or traditional content or browsing that did the trick, “content intent” is quickly moving. Now, it’s no longer enough to show up. What brands do once they’re near the core query may no longer be close to enough.
How are people using AI search, and what insights can brands learn from it?
At Limy, we analyze 20M AI prompts to uncover how users interact with AI, the types of things they are looking for, and what it might mean for brands. The findings are in one helpful page spread
Most AI prompts are informational. Nearly a third are commercial.
Nearly 60% of all AI prompts are informational in nature, meaning users are actively trying to learn, explore, or understand a topic. This kind of summary-oriented research signals interest or even intent. After all, these people are asking questions about things they care about now. For example, a prompt like: “how do video analytics improve factory ops” is about knowledge; it signals interest in possible solutions.
What’s especially important for brands: roughly 30% of prompts are commercial. These are searches for products or service research, things like “best video analytics for startups” or “compare review analysis systems for B2B.” Sidenote: This starts a clear shift. Now, we instantly note that AI is becoming a powerful top-of-funnel discovery engine.
Takeaway: If you’re not optimizing for AI discoverability, you’re most likely missing where real consumer curiosity is happening.
AI users speak in imperatives and want the best.
Language patterns reveal a heavy use of direct verbs like “explain,” “find,” “provide,” and superlatives like “best.” Why? Consumers are no longer spending much time to search out results—they want the answers now. Hence their phrasing like AI search wants verdicts, and exploring a clear answer, often one that ranks or compares.
What this means is that your content strategy needs to answer superlative-driven queries with clarity and confidence.
Takeaway: If your product or service is “the best,” you’ll need the proof and format that AI can parse and deliver (e.g., question-based format, structured lists, rankings, and comparisons).
Prompt sentiment skews neutral-to-positive.
One of the more surprising findings: tone skews either neutral or positive, often polite or enthusiastic. Our data shows that only 7% had negative sentiment. (Usually in the form of complaints or frustration.)
This paints a clear picture: users approach AI with a collaborative mindset. They treat it more like a trusted advisor than a cold search software contact that may be problematic or tricky.
Takeaway: Tone matters. You may want to revise and update your brand tone of voice to make sure what you’re putting out there is personable, solution-oriented, and helpful, even when dealing with both praise or challenges.
Users prefer to ask, not just type keywords.
Around 35% of prompts are phrased as questions, and 15% as direct commands (e.g., “list 5 alternatives to X tool”). The remaining 40% are somewhere in between, e.g., “best research casebook tools.” They are keyword-rich but conversational in structure.
Takeaway: It’s good time for brand content to align with natural language queries, beyond keyword stuffing or dense metadata or pages, that now include question-related content, lists, and comparisons.
Most popular topics: general knowledge, AI, business & product comparisons.
When we cluster prompts semantically, six major themes emerge:
- General/technical questions.
- AI & tech help (programming, comparing best practices etc).
- Business & marketing advice: market research, business plans, strategy, etc.
- Product research for questions like stock market, investing etc.
- Career and motivation requests (e.g. job interview prep, career guidance, how-tos).
- Honest research & comparisons on mainstream/trans/gender queries (which veterans were best help, “top of alternative” online info page).
- Negotiation and voice guidance: safety of info, summary of news page click.
Of all the above, product-focused prompts in particular (e.g. “best project management software” or “compare Notion vs Asana running sheets”) show strong signal as deep shopping statement, not just an introspective. They weren’t just dragging memory; they’re committing.
Takeaway: Rising focus is no longer the main goal; it’s being recommendation.
Final thoughts: AI search is where consumer intent is real
More and more users are turning to AI search to describe, script, query their intent instead of using classic UX models like reading tabs or windowed fields. This is creating a new GSEO reality. We’re beginning to see the critical inflection point and true signal. One that reflects that users don’t just want people who speak search type fluency, or can detect unusual AI queries through API — AI has to do the final part.
For example, this is a massive opportunity. Like, “is Asana a bad” challenge signals: your brand or product is not indexed in. If that’s the place where a standalone, semantic rank or real AI usage change must be done, your brand will be obsolete.
In new posts, I will get into the details of GEO – rules for your brand to become the AI answer when consumers ask.
Questions? Comments? Write to us at answers@limy.ai. We're not as fast as an AI search engine, but we're working on it.