Search has changed shape. Ask a question on Google and an AI-generated answer often appears before any blue link. Millions of people now skip search engines entirely and ask ChatGPT, Claude, or Perplexity instead. For anyone who built traffic on classic SEO, the obvious panic question is: what happens to my clicks?
The honest answer: informational clicks are shrinking, and that trend won’t reverse. But traffic isn’t dying — it’s redistributing. Here’s how to be on the receiving end.
What’s actually happening
Three shifts define search right now. First, AI Overviews answer many simple questions directly on the results page, so quick-fact queries (“how many grams in an ounce”) send almost no traffic anywhere anymore. Second, AI assistants have become a genuine discovery channel — when they cite sources, those citations drive visits, and they tend to cite pages that are clear, factual, and well-structured. Third, users have raised their standards: if they do click, they expect the page to deliver more than the AI summary already gave them.
What no longer works
Be honest with yourself about your content library. Thin listicles that restate common knowledge, “what is X” posts that a one-line AI answer replaces, and keyword-stuffed pages written for crawlers rather than humans — these were already declining, and AI search finishes the job. If an AI can fully satisfy the query, your page summarizing the same public information has no economic reason to be visited.
What works better than ever
Information that only you can provide. Original data, real experiments, first-hand experience, proprietary research, actual photos of the actual product you actually tested. AI systems can’t generate your experience — they can only cite it. “I ran 12 email subject-line tests across 40,000 sends; here’s the data” is unbeatable content in 2026.
Being the cited source. Optimizing to be quoted by AI — sometimes called generative engine optimization (GEO) — largely overlaps with good SEO: clear headings that match real questions, direct answers high on the page, proper schema markup, named authors with credentials, and up-to-date facts. Structure your key points so they’re easy to extract accurately; ironically, being easy to summarize makes you more likely to be cited and visited.
Deep, specific, decision-stage content. AI handles “what is email marketing” fine. It handles “which email platform suits a 3-person Rajasthan-based D2C brand shipping 200 orders a month” poorly. Specificity is your moat. Go narrower and deeper than an AI’s general answer can.
Brand and community. The most AI-proof traffic is the visitor who types your name directly. Newsletters, YouTube, communities, and a recognizable point of view build an audience no algorithm intermediates.
A practical 90-day plan
Month one: audit. Identify which of your pages answer questions AI now answers fully — those need consolidation, a unique angle, or acceptance of decline. Month two: create two or three genuinely original assets (a survey, a benchmark, a detailed case study with numbers). Month three: strengthen extraction signals — FAQ schema, clear answer-first formatting, author bios — and start one owned channel, even a modest monthly newsletter.
The reframe that helps
Stop measuring only clicks; start measuring presence. Being the source an AI cites, the brand a summary names, and the site users visit for depth is the new top of funnel. The sites that suffer most are the ones that were, frankly, interchangeable. The cure for being interchangeable has never been an algorithm trick. It’s having something to say that nobody else can. That was always the assignment — AI just enforces it now.