You cannot improve what you cannot see: baselining your AI-search visibility

The channel that now decides enterprise shortlists is the one most firms do not measure at all. Zero-click searches reached 69% in the year after AI Overviews launched, and organic click-through on queries with an Overview fell 61% between mid-2024 and late 2025. But in one Seer Interactive case study, traffic from ChatGPT and Perplexity converted at 15.9% and 10.5%, against 1.76% for Google organic, and AI-referred sessions grew 527% year over year. The visibility moved; the measurement did not follow.

Ranking #1 on Google is no longer the same as being the answer. If you are not tracking citations in ChatGPT, Perplexity and Gemini, you have no read on your best-converting channel.

Why your SEO dashboard is now blind

The share of Google AI Overview citations that also rank in Google's top ten has fallen from 76% in mid-2025 to 38%, according to Ahrefs. Your rank tracker measures a surface that AI answers increasingly ignore. Worse, the platforms diverge from each other: research drawing on 680 million AI citations estimates only 11% of domains are cited by both ChatGPT and Perplexity. There is no single ranking to chase; there is per-engine visibility to measure.

What a real GEO baseline looks like

You do not need a new analytics religion. You need a repeatable measurement of whether the engines that frame buying decisions treat you as a source, tracked over time against your competitors.

• A fixed set of 20–30 buyer-intent prompts run monthly through ChatGPT, Perplexity, Gemini and Claude.
• A log of where you and each competitor are cited, by engine, so you optimize per platform (Perplexity leans on community sources such as Reddit; ChatGPT favors encyclopedic sources such as Wikipedia and Bing's top results).
• Referral analytics segmented by AI source, so you can see the conversion premium AI traffic carries.
• Bing Webmaster coverage confirmed: ChatGPT web search draws heavily on Bing, so Bing indexing is the price of entry.

The HOWF position

We stand up the baseline before we touch a line of content: the prompt set, the per-engine citation log, the AI-segmented referral view. Only then do we optimize (structure, entity graph, and third-party authority) and re-measure monthly. GEO without measurement is a story you tell yourself. With it, being the answer becomes a number you can move.