Users need access to detailed metadata about AI search conversations, including prompt words, number of messages per conversation, conversation duration, and how many conversations lead to external website visits. This data is crucial for understanding and optimizing for natural language search.
AI Search query research is messy and complex and it should be. But don’t be a Cranky McCrankerface and decide that it has no value because the data isn’t perfect. Nobody likes Mr. Crankers. Listen. For years, SEO revolved around keyword search volume. People typed two, three, maybe four words into a box, and we optimized around those head terms. It worked. Cool. Now, more people are using full sentences and questions. They’re talking to their phones (gasp). They’re using conversational platforms (gasp). The shift to natural language means those old head terms start to lose their functional value. In the context of AI Search, query fan out infuses related searches. So it’s not just about those big monolithic terms anyway. But we’re resourceful marketers. We need ways to generalize topics while still gaining insight into search volume. That’s where the AI Search tools like Profound are starting to get interesting. I’ve been playing with the new features (not sponsored, but we like the peeps over there): • Profound’s prompt volume data shows real behavioral signals from millions of opted-in users. It’s in the FAQs. • It lets you break down prompts by persona, intent, platform, age, gender, and income. • It helps you see search intent (and segmented search intent). Are you writing a short story about running shoes or buying running shoes? What’s commercial intent how much of search is not transactional at all (yet - I haven’t bought shoes via ChatGPT yet). Take the running shoes example. ProFound shows that 54% of prompts aren’t even commercial. They’re informational. Only 27% fall into mid-funnel, consideration-type queries. That’s the gap where brands like Nike, Adidas, and New Balance should be investing in content. That’s how you align an AI search content strategy, with intent awareness, evaluation, purchase, and support, instead of guessing where to spend your time. It’s not perfect data. It’s clickstream data. It’s biased by who opts in. But keyword search volume has never been perfect either. It’s directional insight. And, in my humble (?) opinion, direction is better than darkness. We’re entering a stage where query fan-outs will define visibility. Each broad query generates several related prompts expanding into every modifier and context possible. Your content needs to live across those journeys. AI search won’t reward a single “perfect” keyword. It will reward coverage of nuanced, interconnected intents. Check out the video. If you want to see what AI search volume looks like in Profound for a keyword, drop it in the comments. I’ll share some screenshots.