If you’ve sat through a marketing planning meeting in 2026, you’ve probably heard some version of this pitch: AI search is the future, so you need to invest in “GEO” — generative engine optimization — to show up inside ChatGPT, Gemini, and AI Overviews. It’s a reasonable-sounding idea. But three independent analyses published within the same week in October 2026, each using real website analytics rather than platform-reported figures, arrived at a strikingly consistent and much less dramatic picture of how much traffic AI assistants actually send today.
Here’s what the data shows, what it doesn’t prove, and how to think about your own 2027 budget in light of it.
Haynes analyzed 12 months of Google Analytics 4 data across 100 websites of varying size, totaling more than 145 million sessions. She found that sessions arriving from ChatGPT amounted to just 0.62% of the volume driven by Google organic search. Interestingly, ChatGPT visitors converted at a notably higher rate — 13.8%, versus 3.02% for Google organic visitors. But because the volume gap is so large, Google organic still accounted for 96.5% of total conversions across the channels she analyzed, and drove roughly 27 times more conversions than all AI assistants combined.
This analysis (sample size not disclosed in available reporting) found that organic search delivered about 39 times more traffic than AI assistants. For every 100 conversions organic search produced, AI referral traffic produced about 3.2. AI referral traffic converted worse than paid search (at roughly 0.67 times the rate) but better than direct traffic (about 2.48 times).
Looking across 12 sites in different sectors, Indig could not detect a measurable increase in direct traffic to long-tail pages that would suggest AI assistants are quietly driving unattributed (“dark”) traffic. He found no reliable relationship between AI visibility and branded search or direct traffic — while being careful to note that an absence of a detectable effect in his sample isn’t proof that no such effect exists anywhere.
It’s tempting to read “AI sends almost no traffic” and conclude AI search doesn’t matter. That’s too simple, for two reasons.
First, Haynes’s own data shows AI-driven visitors convert at more than four times the rate of Google organic visitors. The traffic volume is small, but the quality — at least in her dataset — is notably higher. That looks less like “AI search doesn’t matter” and more like “AI search currently sends a small number of unusually high-intent visitors.”
Second, there’s a real attribution problem underneath all three studies. Analytics tools can struggle to correctly tag traffic that originates from an AI assistant’s answer, especially when a user copies a link or searches separately after reading an AI summary — this can get misclassified as “direct” traffic. Indig’s study was specifically looking for this effect and couldn’t confirm it, which is useful, honest evidence — but it’s evidence of absence in his 12-site sample, not proof that it doesn’t happen more broadly.
It’s also worth sizing the AI side of this correctly. OpenAI’s own head of ChatGPT search, Adam Fry, said in a November 4, 2026 interview that ChatGPT’s search-like query volume is roughly one-fifth of Google’s, and that ChatGPT’s user base grew from 1 billion to 1.2 billion in the preceding two months. That’s a measure of how much people are using ChatGPT to ask things, not how much referral traffic it sends to websites — a distinction worth holding onto, since the two numbers answer different questions.
Be precise about what’s actually been measured here. All three studies report traffic volume and conversion rates. None of them calculate a return on investment for GEO spending — that would require knowing the cost of the AI-visibility work being evaluated, which none of the three studies address. The framing that AI visibility is “not worth investing in” is a reasonable interpretation of the traffic numbers, but it’s an editorial conclusion, not something the underlying data directly proves.
It’s also worth being upfront about limits: Haynes’s sample mixes websites of different sizes and industries; SEO Works hasn’t published its sample size; and Indig’s 12-site sample is small. None of the three studies publish an India-specific breakdown.
A few practical conclusions, held carefully against what the data actually shows:
Before shifting your 2027 marketing budget into GEO, audit your own data and measure exactly how much high-intent traffic ChatGPT, Gemini, and Perplexity are driving to your site today.
A 15-minute session?
Take the first step toward future-proof growth:
For a Direct Business Query: Call us at 99034 72194
For a Quick Consultation: WhatsApp our team at 79807 31010
#AISearch #GEO #SEO #MarketingBudget2027 #ContentStrategy #ChatGPTAds #AIOverviews #DigitalMarketing #B2BMarketing #MarketingROI #TejomDigital
New AI shopping adoption statistics show India has the highest consumer trust in AI shopping tools of any market McKinsey surveyed. Here’s what the data actually says, what’s still unverified, and what it means for how Indian brands should prepare.
Google’s dynamic search ads to AI Max migration deadline has moved from September 2026 to February 2027 after advertiser pushback. Here’s exactly what already changed, what’s still coming, and the account checks to run before the forced cutover.
Google Search Profiles are now live in India. This free feature pulls your social accounts, posts, and links into one discoverable page on Google Search and Discover — but only if you meet Google’s follower threshold. Here’s exactly who qualifies, how to claim your profile, and why it’s worth a founder’s attention.
A new academic study measuring real advertiser and website data finds ChatGPT ads revenue per visit runs well behind Google Search ads. Here’s exactly what the research found, what it doesn’t prove yet, and how it should inform your 2027 ad-budget decisions.
Google has quietly confirmed that Performance Max can now filter out brand and navigational searches and send them to Search campaigns instead. Here’s what the Performance Max brand query filtering change actually does, why it matters for your budget and reporting, and exactly what to check in your account this week.
Diwali 2026 just wrapped. For brands with any reach into Bihar, Jharkhand, or eastern UP communities, Chhath Puja — falling November 13-16 — is a second, genuinely separate festive moment. Here’s a chhath puja 2026 marketing strategy built on what brands actually did last year, not assumptions.
Google Ads can now pull your own videos and landing-page images into live Performance Max ads automatically — and from December 4, a second default takes effect too. Here’s exactly what changed, why it matters for brand control, and how to opt out if you want to keep it.
Google and NielsenIQ just announced a partnership to build the data backbone behind AI shopping agents. The Google NielsenIQ agentic commerce deal isn’t a consumer feature — it’s infrastructure. Here’s what it actually says, and what’s still just a vendor claim.
