Summary: In March 2025, Adobe Analytics found AI-referred traffic to US retail sites converted 38% worse than non-AI traffic. In March 2026, the same measurement found it converted 42% better. That's the fastest reversal of a channel-quality metric most e-commerce teams will see in their careers — and it changes the budget question from whether to fund AI visibility to how much, taken from where.
The honest version of the argument
Most content on this topic argues that AI search is replacing SEO and you should move budget accordingly. That argument overstates its case, and D2C operators can smell it.
Here's the more defensible position: AI search and organic search are now two channels with very different volume and very different quality, and the correct allocation depends on which of those two things is currently your constraint.
Both halves of that sentence are supported by published data. Let's take them in order.
What the volume data says
Search itself is leaking clicks, steadily and measurably.
- SparkToro, using Datos clickstream data: 68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% in 2024 — a 7.56-point rise in two years. The share of searches producing at least one click fell 9.51 percentage points over the same period, a 22.9% relative decline.
- Pew Research Center, in a behavioural study of 900 US adults across 68,879 real Google searches, found users clicked a traditional result 8% of the time when an AI Overview was present, versus 15% when it wasn't — a 47% relative drop.
Meanwhile AI referral volume is climbing fast from a small base. Adobe Analytics — drawing on more than a trillion visits to US retail sites — reported AI-referred traffic to US retailers grew 393% year-on-year in Q1 2026.
That 393% is the number every vendor quotes. The number they leave out is the denominator. AI referral traffic remains an order of magnitude smaller than Google organic for most retailers. A 393% increase on a small base is still a small base.
So: search volume is eroding, and AI volume is growing fast but hasn't replaced it. Neither channel can be abandoned. That's the volume picture, and on volume alone the answer would be "keep doing SEO."
What the quality data says
This is where it gets interesting, and where the case for reallocating actually lives.
Adobe's comparison of AI-referred versus non-AI traffic to US retail sites found:
| Metric | AI-referred vs non-AI traffic |
|---|---|
| Conversion rate (March 2026) | 42% better |
| Conversion rate (March 2025) | 38% worse |
| Time on site | +53% |
| Pages per visit | +23% |
| Revenue per visit | +37% |
| Engagement rate | +12% |
Read the first two rows together. Twelve months earlier, the same measurement said the opposite. This is not a stable, well-understood channel — it is a channel that materially changed character inside a year, and any strategy built on 2025 data about it is wrong.
Corroborating evidence from a different sector and methodology: Ahrefs reported that 0.5% of its traffic originating from AI search drove 12.1% of all signups over a 30-day period — roughly a 23x conversion rate differential.
The mechanism is intuitive. Someone arriving from an AI answer has already had alternatives compared for them, objections partially handled, and a specific reason to click. They arrive further down the funnel than a search visitor does. The visit is worth more because more of the selling already happened.
Caveat worth stating plainly: "conversion rate" is not defined identically across these studies. Adobe measures retail transactions; Ahrefs measures product signups. They point the same direction, but they are not the same measurement, and you should not average them.
The decision framework
Two questions decide your allocation. Answer them honestly.
Question 1: Is your constraint volume or efficiency?
If your constraint is volume — you need more people at the top of the funnel, your CAC is acceptable, you're not saturated — organic search still delivers far more absolute sessions than AI referral. Keep the SEO investment. Add AI visibility as a measured pilot, not a replacement.
If your constraint is efficiency — you have traffic but poor conversion, blended CAC is rising, paid is saturating — the Adobe numbers argue for reallocating toward AI visibility. You would be buying a smaller volume of substantially better-qualified visits.
Most D2C brands past their first ₹5 crore of revenue are efficiency-constrained, not volume-constrained. That's the honest reason this shift matters commercially.
Question 2: Do you actually know your current AI visibility?
Almost nobody does, and this is the more important question.
You cannot make a rational allocation decision between two channels when you have a measurement for one and a vibe for the other. You know your organic sessions, ranking positions and organic revenue to two decimal places. For AI you probably know that a colleague asked ChatGPT once and the brand came up.
Before moving any budget, get a baseline. Not a single query — a proper one. That means a prompt set covering the way customers actually ask, replayed enough times per engine to produce a number with a confidence interval rather than an anecdote. The reasoning behind replay counts and confidence bands is published in our methodology.
Where the two disciplines overlap — and where they don't
A persistent claim in the SEO industry is that good SEO automatically produces AI visibility. The published evidence does not support treating this as reliable.
| Study | AI-cited URLs also ranking in Google top 10 |
|---|---|
| Ahrefs (2025) | 76.1% |
| BrightEdge | ~17% |
| Search Atlas (18,000+ queries) | ~12% |
A 12%-to-76% spread means the honest answer is it depends on your category, your engines and your query mix — which is another way of saying it's an empirical question about your specific brand.
More striking: Semrush found Google's own AI Overviews and AI Mode share only 13.7% URL overlap with each other. Two surfaces from the same company, citing largely different sources. If Google's two AI products don't agree on what to cite, "rank well and you'll be fine" is not a strategy.
We compare the two disciplines in detail in AEO vs SEO. The short version: they are complementary, not substitutes, and the overlap is smaller than the SEO industry would prefer.
What actually moves the number for D2C
Assuming you've decided to fund this, the work that matters for a consumer brand differs from the B2B playbook.
- Product data as structured data.
Product,Offer,AggregateRatingandReviewschema on every PDP, with current price and availability. Engines answering "where can I buy X under ₹3,000" need machine-readable price and stock. Adobe's own analysis flagged that retail sites lag in being machine-readable — a gap that's unusually cheap to close. - Comparison and alternatives content. Consumer AI prompts skew heavily comparative: "X vs Y," "alternatives to [brand]," "best [category] for [constraint]." If you haven't published the comparison, a competitor or an affiliate site has, and the engine reads their framing of you.
- Third-party corroboration. Large-scale citation analysis by Peec AI (roughly 30 million citations, March 2026) found Reddit the most-cited domain across ChatGPT, Google AI Mode, Gemini, Perplexity and AI Overviews. For consumer categories, community discussion and review platforms carry disproportionate weight relative to your owned content. Mapping which specific domains feed the prompts you lose is what citation intelligence is for.
- Fact accuracy monitoring. Stale pricing in an AI answer is a direct revenue leak — customers arrive expecting a price you no longer offer, or don't arrive because they were quoted one you never charged. Price accuracy in AI answers deserves the same monitoring discipline you already apply to marketplace listings.
- Attribution, so this survives the next budget review. The reason AI visibility work gets cut is that nobody connected it to revenue. Tying citation movement to sessions, conversions and revenue via GA4 and your commerce stack is the difference between a funded programme and an experiment. That's the job of revenue attribution.
A proposed allocation
Not a rule — a defensible starting point, to be revised the moment you have your own data.
| Stage | Suggested split | Trigger to move on |
|---|---|---|
| Baseline (month 1) | 100% existing SEO + a fixed-cost audit | You have a measured AI visibility score with a confidence band |
| Pilot (months 2–4) | ~85% SEO / ~15% AI visibility | Measurable citation-share movement in at least two engines |
| Scale (months 5+) | Reallocate against measured revenue per channel | Ongoing |
The important discipline is in the trigger column. Move budget on evidence, not on a conference talk. If the pilot doesn't move citation share in two engines within three measurement cycles, either the execution or the thesis is wrong for your category — and you should find out for a small number rather than a large one.
The one-paragraph version
Organic search still delivers more volume; AI referral delivers better-qualified visits, and the quality gap flipped from negative to strongly positive within twelve months. Fund both. Measure the one you currently can't see before you move money toward it. Be sceptical of anyone — us included — quoting a confident single number about AI citation behaviour without publishing their method.
If you want the baseline: run a free AI visibility audit, or look at the labelled sample report first. The D2C and e-commerce playbook goes deeper on the retail-specific prompt clusters.
Related reading
- What Is Generative Engine Optimization (GEO)? The Complete 2026 Guide
- AEO vs SEO — how the disciplines relate
- How B2B SaaS Buyers Use AI to Build Shortlists
- 98.8% of Multi-Location Brands Are Invisible to ChatGPT
- GEO for Financial Services
Sources
All figures are from third-party published research. No CiteRank client data appears in this article.
- Adobe Analytics / Adobe Digital Insights, Q1–Q2 2026 AI Traffic Report (based on 1 trillion+ visits to US retail sites and a survey of 5,000+ US respondents): 393% YoY AI-referred traffic growth; conversion, engagement, time-on-site, pages-per-visit and revenue-per-visit comparisons; March 2025 vs March 2026 reversal
- SparkToro (Datos clickstream data), 2026: US zero-click share of 68.01% Jan–Apr 2026 vs 60.45% in 2024
- Pew Research Center, behavioural study of 900 US adults / 68,879 Google searches: 8% vs 15% click-through with and without AI Overviews
- Ahrefs, June 2025: 0.5% of AI-search traffic driving 12.1% of signups; 2025 AI Overviews citation-overlap analysis
- BrightEdge; Search Atlas (18,000+ queries): competing citation-overlap findings
- Semrush: AI Overviews vs AI Mode URL overlap of 13.7%
- Peec AI, March 2026: analysis of ~30 million citations
Published 7 August 2026. AI search measurement is revised frequently; check for updated versions of each study before quoting these figures.