Search used to ask people to shrink their complicated lives into a few keywords. AI search is reversing that. Longer, more personal prompts now carry details we once handed to a travel agent: who is travelling, what children need, which trade-offs matter. For marketers, that makes every prompt a slice of customer research.
The shift is not that customers became more demanding. They always had these constraints. AI search simply made it practical to say them out loud.
From compressed keywords to full briefs
For years, a family holiday query looked like “best family hotel Tenerife.” That phrase hides budget, room layout, food preferences, pool requirements and a lot more. An AI assistant can now receive a full brief: two adults, two young children, a separate sleeping area within the same room, good food, a rental car, swimming pools and a heated kids’ pool.
That level of detail used to emerge later, in a conversation with a salesperson or travel agent. Now it arrives at the discovery stage. Brands therefore need content that answers specific situations, not just broad categories.
Why “best” is becoming situational
Consider two people asking an AI assistant the same question about running shoes. One has previously discussed a knee injury; the other has not. The words in the prompt may be identical, but the assistant’s memory can produce different recommendations. The injury does not need to be repeated in the prompt to shape the answer.
This is a big change. “Best running shoes” is no longer one competition. It becomes many competitions: best for a first marathon, best for a recovering knee, best for comfort over speed, best for someone willing to spend up to $150. There is no single right answer, only the right answer for a person in a situation.
Prompt tracking: useful but incomplete
Marketers naturally want to track what ChatGPT or other assistants say about their brand. That is worth doing, but standardized prompts can only show one version of a personalised answer. If two users get different responses from the same words, a single test prompt is not the whole truth.
Look beyond prompt tracking to first-party signals:
- Which pages receive traffic from AI platforms.
- Which increasingly specific content attracts visits.
- What those patterns reveal about the situations where your brand is surfaced.
That gives a fuller view than any single prompt check.
Use prompts to shape proposition, not just SEO
The real value is upstream. Search teams can now feed richer customer questions into product, pricing and positioning decisions. Ask who you are genuinely brilliant for, under what circumstances you are the right choice, what evidence proves it, and where there is an unmet combination of needs nobody serves. Those are business strategy questions, not SEO tasks.
Internal teams are a second source. Salespeople, customer service and store staff hear the oddly specific questions that never make it to the website. Marketers should feed that knowledge back into content and search strategy. As AI gets better at understanding individuals, brands need the same understanding.
A simple frame to apply: move from intent — what someone typed — to situation — the context behind it — to evidence — the proof that your brand fits that context. This keeps content and targeting tied to real human decision-making, not just query volume.
Search has shifted from keywords to prompts, but prompts are not the final destination. The brands that win will be the ones that treat prompts as customer research and act on the human detail inside them.
Source: MarTech
