What upstream intent is, precisely
Downstream intent is a query: the buyer knows the category, names it, and compares options, the territory every marketer already fights over. Upstream intent is a situation: the buyer knows the pain but not necessarily the category, and describes circumstances, we just doubled headcount and onboarding is chaos, my parents are aging and live far away, our rent went up and margins are gone. Classic search forced people to compress situations into keywords, destroying the signal; conversational AI preserves it. When an assistant hears the situation, it infers the need and recommends solutions, which means shortlists now form at the situation stage, one full step before any keyword tool can see demand. That inversion, recommendation before query, is the thesis we laid out in what you are being taught about AI visibility is wrong, and upstream intent is its raw material.
Why upstream moments are cheaper to win
Downstream keywords are contested precisely because they are measurable: everyone's tools see the same volume numbers, so everyone bids on the same terms. Upstream moments barely register in keyword tools, they have no stable phrasing, so almost nobody optimizes for them, and the brands that do compete with absence rather than incumbents. They also convert differently: a brand recommended inside the buyer's own described situation arrives with borrowed trust and situational fit already established, which is why assistant-referred buyers behave like referrals. And they compound defensively: once an assistant associates your brand with a moment, a competitor cannot outbid you for it; they have to out-evidence you, which takes quarters, not budget. The mapping discipline for finding your category's moments is AI presence mapping, formalized in CAPM©.
The capture method, end to end
Capturing upstream intent runs in four moves. Discover: mine the situations, customer interviews about what was happening the week they started looking, community threads where buyers narrate circumstances, and the assistants themselves, describe a situation and see what gets inferred. Cover: publish content that speaks to the situation in the buyer's own language, before category vocabulary, the page that says our onboarding broke when we doubled headcount outranks the one that says workflow automation solutions in the moments that matter. Evidence: earn the third-party trust, reviews and threads mentioning your brand in situational language, that lets assistants confidently attach you to the moment. Measure: track a basket of situational questions monthly across the four assistants, per AI search tracking, and tie movement to revenue through the chain in tying GEO results to revenue. Aethon runs this loop as its core product; the free shows you today's upstream coverage in a minute.