Google, Yahoo, Perplexity, and other search companies are working on understanding user intent to decide whether to show generative AI answers or just links
Google had already tied generative AI to query context in Search advertising, through its plan for AI-assisted ad relevance and campaign creation. The reported intent-routing work extends that contextual judgment from ads into the choice of search interface itself.
Perplexity’s paid answer-engine offering had made the competitive premise visible: AI-generated responses could provide an alternative to conventional result pages. Perplexity Pro’s early search experience framed the pressure on incumbent search providers to distinguish when an answer is more useful than a list of sources.
First-order effects
Google, Yahoo, Perplexity, and peers must classify queries before rendering results, making the search experience conditional on inferred user intent rather than a single default format.
Users may receive either generated responses or link-led results for similar topics depending on what the system determines they are trying to accomplish.
Second-order effects
Ranking, retrieval, and answer-generation teams become more tightly coupled: a weak intent decision can send users to an expensive or unsuitable AI response, or deny them a direct answer when one is useful.
Publishers and marketers face less predictable referral exposure as some queries resolve within an answer surface; later interest in tools for appearing in AI search summaries reflects the incentive to optimize for that surface.
Third-order effects
If intent-based routing becomes standard, search competition shifts from ranking pages alone toward deciding which queries merit synthesis, citations, and onward clicks.
The model could create a more segmented search economy, with answer-oriented queries concentrating attention inside search products while link-oriented queries remain the principal discovery path for the open web.
The trend: Search is evolving into an intent-routed interface that selectively combines generative answers with conventional web discovery.
Alternate translation: Tech firms trying to find way to give you “AI” generated responses that help clients rather than you, but must find way to make it harder for you to know it was AI but still not violate FTC rules about labelling products. [embedded post]
How about you just ask me? A little button, a check box? You know, asking is WAY CHEAPER than having teams “work on understanding” what I want. [embedded post]
Bing already does this based on whether you type in keywords (when you might get genAI chunks with the links) or a question, which it goes ahead and answers so maybe intent isn't that hard? [embedded post]
So I've read a lot of people talking about perplexity AI, and I feel like I'm going crazy. $520m valuation on $10m annual revenue??? For a $20 a month search engine that uses AI? https://www.nytimes.com/... [image]
Startups are hiring influencers to peddle bullshit. This is a new kind of misinformation to peddle snake oil AI products with no backing. When you dig in, you see the growth rate has been hovering around 8.5%, with leading and lagging indicators for March showing a -16% drop. [im…
Shame on me for trying to get a story out during earthquake Twitter, but if you care about search here's a one for you: https://www.bigtechnology.com/ ...