ChatGPT recommendations are creating demand that analytics often credits to search
Similarweb’s study shows why AI visibility is not only a referral problem: ChatGPT can shape brand choice first, while the eventual visit often arrives through search, direct navigation or another measurable channel.
Editor’s Note by Dageno: Dageno AI’s editorial team structurally distilled Search Engine Land’s article “ChatGPT recommendations drive more brand website visits: Study” to help GEO and AI visibility teams connect AI recommendation exposure, downstream search demand and attribution blind spots.
Understand the article in 60 seconds
- Similarweb found that users who received a ChatGPT brand recommendation were 2.5 times more likely on average to visit that brand than a direct competitor within seven days.
- The study tracked U.S. desktop journeys in finance, travel and beauty, excluding users who had visited the recommended brand in the previous four weeks or named it in the prompt.
- AI-influenced visits often looked like search traffic: 55.9% came through search versus 40.4% of non-AI-influenced visits, while direct traffic was lower at 19.9% versus 38.8%.
- Recommended users engaged more deeply, averaging 12 pages and 11.8 minutes on site compared with 6.5 pages and 5.6 minutes for other visitors.
AI recommendations create demand before analytics can name the source
The article’s core point is that ChatGPT can affect brand choice even when the visit does not show up as AI referral traffic. The influence happens upstream: the user asks a category or industry question, receives a brand recommendation, then later searches or navigates to the brand site.
For GEO teams, this means AI visibility is an answer-level demand signal. If reporting only looks for traffic from chat.openai.com, it will miss much of the impact because the user may return through Google, Bing, direct navigation or another channel.
Similarweb’s design isolates recommendation influence, not ordinary brand loyalty
Similarweb tracked U.S. desktop users who asked ChatGPT industry-relevant questions, received a specific brand recommendation and then visited either the recommended brand or a direct competitor within seven days. The study covered finance, travel and beauty brand pairs.
The exclusions matter. Users who had visited the brand in the prior four weeks were removed, and prompts that named the brand were excluded. That makes the signal more useful for GEO because it focuses on category discovery rather than existing customer behavior.
The lift appeared across finance, travel and beauty rather than one isolated vertical
The pattern was visible in multiple competitive pairs. After an American Express recommendation, 7.2% of users visited American Express versus 3.1% visiting Capital One; after a Capital One recommendation, 14.2% visited Capital One versus 3.8% visiting American Express.
Travel and beauty showed the same logic: Skyscanner and Kayak recommendations shifted visits toward the named brand, and Sephora or Ulta recommendations produced stronger downstream visits than the competitor baseline.
Search becomes the capture channel for demand created by AI
The most important measurement detail is that AI-influenced users often arrived through search. Similarweb found 55.9% of AI-influenced visits came through search, compared with 40.4% among other visits. Direct traffic moved in the opposite direction, at 19.9% versus 38.8%.
That means search reports may be capturing demand that ChatGPT helped create. A rise in branded queries after AI recommendation exposure should not automatically be credited only to SEO or paid search.
AI-recommended visitors arrive with narrower consideration sets
AI-influenced visitors viewed 12 pages and spent 11.8 minutes on site on average. Non-AI-influenced visitors viewed 6.5 pages and spent 5.6 minutes. The likely reason is that the AI conversation already helped users narrow choices before they clicked or searched.
For growth teams, that makes AI visibility not only a top-of-funnel metric. It can change visitor quality, comparison behavior and downstream conversion probability.
The study turns AI visibility into a measurable demand path
The facts in the Similarweb study can be converted into diagnostic checks for brand teams.
2.5x visit likelihood
Recommendation exposure can redirect brand demand before any website session is recorded.
Seven-day window
The user journey is delayed, so same-session attribution understates AI influence.
Three verticals
Finance, travel and beauty suggest the effect is strongest where users compare options and trust matters.
Deeper on-site engagement
AI-recommended visitors may arrive with stronger intent and a shorter consideration list.
Where the downstream signal can disappear in reporting
The measurement problem is not that AI has no impact; it is that the impact can be reclassified by analytics systems.
| Observed channel | Hidden AI role | What teams should test |
|---|---|---|
| Organic search | ChatGPT may have created the brand preference before the search query. | Compare branded search lift against prompt-level recommendation exposure. |
| Paid search | The click may capture demand that AI generated earlier. | Segment brand campaigns after AI visibility changes and monitor CPC, CTR and CVR. |
| Direct traffic | Some users navigate after AI-driven consideration without referral data. | Survey new users and compare direct visits after high-volume prompt wins. |
Teams need to connect AI visibility, evidence and attribution owners
The article points to an operating problem: AI visibility is not owned by one team. The response has to assign who changes content, who measures demand, who manages trust and who interprets competitor leakage.
| Owner | What changes | What to do next | Signal to monitor |
|---|---|---|---|
| GEO team | Answer-share becomes a demand generation metric. | Track recommendations across non-branded category, comparison and problem prompts. | Brand recommendation rate and competitor answer-share. |
| SEO team | Search may capture demand initiated inside ChatGPT. | Separate branded query growth from ranking-only explanations. | Branded impressions, query variants and AI-exposed prompt families. |
| Analytics team | Referral reports undercount influence. | Create assisted visibility models linking AI exposure to search, direct and conversion quality. | Search-mediated AI lift and engagement deltas. |
| Content strategy team | AI-recommended visitors need decisive proof after arrival. | Build comparison, trust, pricing, FAQ and use-case pages for recommended journeys. | Pages per session, time on site and assisted conversion. |
| Brand team | Recommendation quality shapes preference before the site visit. | Manage third-party descriptions, reviews and category positioning that ChatGPT may rely on. | Sentiment, entity consistency and third-party source presence. |
Measurement should expose prompt visibility, search mediation and competitor leakage
The useful metric is the one that tells a team what to fix next. These signals connect answer visibility to content updates, brand governance and growth attribution.
| Signal | Why it matters | Owner | Response trigger |
|---|---|---|---|
| Prompt recommendation rate | It shows whether ChatGPT is creating demand for the brand or for competitors. | GEO team | Recommendation share drops on high-intent category prompts. |
| Search-mediated visit share | It catches AI influence that appears as search traffic. | Analytics | Search share rises after AI wins but AI referral remains low. |
| Branded query lift | It indicates that AI exposure may be prompting later search behavior. | SEO | New branded modifiers appear after prompt visibility changes. |
| Engagement depth | It shows whether recommended users arrive more qualified. | Growth | AI-exposed cohorts view more pages but fail to convert. |
| Competitor follow-through rate | It reveals whether ChatGPT is sending demand to rival sites. | Leadership | Competitors receive stronger downstream visits after recommendations. |
How to connect AI recommendations to downstream demand
Track prompt families first
Measure visibility for category, comparison and problem prompts before looking at referral reports.
Add search-path attribution
Compare branded search, paid search and direct sessions before and after AI recommendation wins.
Build post-recommendation pages
Make sure users who arrive after AI comparison find proof, pricing, alternatives and decision support.
Report competitor leakage
Show leadership where ChatGPT recommends competitors and whether those recommendations convert into site visits.
Turn AI visibility signals into execution.
Dageno AI is a complete execution platform from data monitoring → strategy → content generation → result attribution. It helps teams monitor AI answer surfaces, identify strategy gaps, generate content actions and connect the result to visibility and business impact.
Alexa for Shopping