What the GPT-5 search rollout means for organic traffic.

OpenAI flipped the switch on GPT-5 web-search for all paid tiers two weeks ago, and to free users earlier this week. The rollout was quiet — no big keynote, no countdown — but it's the most consequential change to the search landscape since Google introduced featured snippets. Here's what the implications actually are, ranked by how much teams should be reorganizing around them.

Change 1: Citation traffic, not zero-click traffic

The doomer take is that AI search means "zero clicks" — users get their answer and never visit your site. That's mostly wrong. GPT-5's web search returns cited sources at the bottom of every answer, and our click-through data on early access shows ~12-18% click rates on those citations. Lower than Google organic, higher than featured snippets. The traffic pattern is different but the traffic exists.

Change 2: Brand mentions matter more than rankings

Conventional SEO tracking measures "where does my URL show up in the SERP for query X." AI search asks a different question: "when someone asks about my category, does my brand show up in the answer?" That's a brand-mention problem, not a ranking problem. The teams that pivot their reporting fastest will quietly outperform.

Change 3 (the one most teams aren't bracing for)

The third change is the one nobody's talking about: freshness windows are collapsing. GPT-5 with web search now updates its citation pool in near-real-time for hot topics. Content from yesterday can outrank content from last quarter on time-sensitive queries. That changes content cadence assumptions for every team running a slow editorial calendar.

If your team publishes twice a month and your competitor publishes twice a week, the math has shifted against you in a way that won't show up in Google ranking tools for another quarter. Treat AI search like a separate funnel with its own cadence requirements. Most teams don't yet — that's the opportunity.