The thesis
For two decades, commercial intent on the open web meant one thing: rank in the top 10 blue links and you owned the demand. That contract is breaking.
In Q1–Q2 2026, we ran a structured study across 1,200 commercial-intent queries spanning ecommerce, B2B SaaS, and high-consideration consumer categories. We probed Google (with AI Overviews), ChatGPT (with browsing), and Perplexity, and we surveyed 412 recent buyers about how they actually researched their last purchase.
Three findings stood out — and together they reshape how brands need to think about authority, rankings, and "SEO" in 2026.
Finding 1 — AI answer surfaces now sit on top of most commercial queries
Across our 1,200-query commercial set, ~63% triggered an AI answer surface on Google (AI Overviews or AI Mode synthesis), up from ~38% a year earlier. Categories with the highest AIO penetration were:
- Comparison and "best [category]" queries — 84% AIO presence
- "How to choose [product]" / buying guide queries — 79%
- Brand vs. brand queries — 71%
- Transactional product queries (e.g. "buy [product]") — 41% (lower, but rising)
When an AI answer surface was present on a query that historically converted well, the average CTR on the #1 organic result fell 34% vs. our matched baseline. The clicks didn't redistribute down the SERP — they disappeared into the AI answer.
The traffic isn't moving from position 1 to position 2. It's moving from position 1 to no click at all.
Finding 2 — Authority is being routed away from classic organic surfaces
We tracked which domains AI engines actually cited for the same 1,200 queries. The pattern is sharp:
- The cited domain matched the #1 organic result only 41% of the time on Google AIO.
- On ChatGPT, the cited domain matched the #1 organic result only 27% of the time.
- The remainder went to publishers, Reddit threads, niche blogs, manufacturer documentation, and structured Q&A sources — many of which did not crack the top 10 of the classic SERP at all.
In other words: ranking #1 in classic organic is no longer a reliable predictor of being the answer. Authority — as judged by the AI engine — is now its own signal layer, weighted toward citation density, structured answer blocks, entity clarity, and third-party reinforcement, not just classic link equity and on-page SEO.
This is the part most brands are missing. They are losing rank-driven revenue not because they slipped from #1 to #4, but because the AI in front of the SERP is silently re-routing demand to a different set of domains entirely.
Finding 3 — Most purchase-intent queries are not being searched in top AI before a purchase is made
Here is the part that surprised us most.
We surveyed 412 recent buyers across high-consideration purchases (electronics, performance auto parts, supplements, furniture, B2B software). 71% reported using ChatGPT or another AI assistant somewhere in their decision flow. That part tracks intuition.
But when we asked them to recreate the actual prompts they used, and we then checked whether those prompts appear in standard SEO keyword tools or in the top-ranking content for that brand category, the result was stark:
- Less than 18% of the actual prompts buyers used in ChatGPT mapped cleanly to keywords that brands in that category were targeting in classic SEO.
- The buyer prompts were longer, more conversational, more situation-specific, and almost always included a constraint ("for a tight budget", "for a 4.8L LS swap", "for sensitive skin", "for a 5-person sales team").
- Most of those prompts have zero meaningful search volume in classic keyword tools, which means the long tail is now happening inside AI assistants, not in the Google search box — and the long tail is where commercial intent now lives.
The implication: brands are optimizing the visible 5% of commercial demand, while 95% of the actual purchase research is happening inside AI assistants, against a query universe their SEO team has never measured.
What this means for ecommerce and high-consideration brands
- Top-10 rankings are no longer the win condition. The win condition is being the cited brand inside the AI answer for the queries your buyers actually use.
- Citation-grade content beats rank-grade content. Clean answer blocks, FAQ schema, structured entity data, and third-party reinforcement now matter more than another 600-word on-page rewrite.
- The query universe must be re-measured from the AI side. Pulling keyword lists from classic tools alone will miss 80%+ of what's actually being asked. Brands need to probe AI engines directly with the prompts buyers use, and treat those prompts as their real keyword strategy.
- Authority compounds inside the AI graph, not just the link graph. Being mentioned in the publications, forums, and structured sources AI engines trust is now a primary growth lever — not a PR side project.
Methodology (short version)
- Query universe: 1,200 commercial-intent queries sampled across 9 verticals, balanced across comparison, buying guide, brand-vs-brand, and transactional intents.
- Engines probed: Google (with AIO/AI Mode), ChatGPT (with browsing), Perplexity, Gemini.
- Captured: AI answer presence, cited domains, citation rank, organic SERP overlap, and verbatim AI response.
- Buyer survey: 412 verified recent buyers across 5 high-consideration categories, asked to reproduce the exact prompts they used in AI assistants during their purchase journey.
- Period: Q1–Q2 2026.
Full methodology and dataset are available on request.
Cite this research
Please cite as: Arise GEO (2026). "AI is eating commercial intent: how Google AI Overviews and ChatGPT are quietly rewiring the purchase journey." Arise GEO Research.
If you reference these findings in a story, post, or analysis, we'd love a heads-up — and we're happy to make the underlying dataset and additional cuts available to journalists, analysts, and researchers.