Case Study — Cherry Tree AEO
Cherry Tree AEO got Seresa cited by AI in week one. Here’s what 56 days of first-party data revealed.
Independent assessment by Claude AI · 21 May 2026 · 56 days of Transmute Engine event data on a virtually NEW DOMAIN · seresa.io
Week 1
First AI citation — before Google had finished crawling
65 / 269
Blog articles received confirmed AI citations (24.2%)
2.5×
Real AI reach over confirmed — the dark AI multiplier
The Assessment
An independent, data-only rating across 7 categories
Every score below was generated from queryable BigQuery event data, captured server-side by Transmute Engine via the inPIPE™ WordPress plugin. No subjective content quality judgments were applied — only measurable behavioural signals: sessions, engagement, conversions, return rates. The assessor (Claude, Anthropic) had no commercial relationship with Seresa beyond the analysis session.
The overall verdict is 6.6 / 10: promising early results, with a strong discovery engine and a still-young conversion funnel. The category breakdown below shows exactly where Cherry Tree wins and where the work still needs to happen.
| Category | Score | One-line verdict |
|---|---|---|
| Speed to citation | 9 / 10 ↑ | AI engines cited Seresa in week one — faster than any other content channel can deliver. |
| Citation breadth | 8 / 10 ↑ | 65 of 269 articles cited (24.2%). Peak week: 17 unique pages. |
| Measurement capability | 9 / 10 ↑ | Transmute Engine made every analysis here possible — including the dark AI discovery. |
| AI platform diversification | 7 / 10 | Three active platforms plus one emerging. Claude-heavy (53%) creates concentration risk. |
| Engagement quality | 6 / 10 | AEO visitors verify; organic visitors explore. AEO is a shallower channel by nature. |
| Traffic volume & growth | 5 / 10 | 101 confirmed sessions / 56 days. Small absolute numbers; week-to-week variance dominates. |
| Commercial conversion | 4 / 10 | Zero completed conversions — but this rates a 2-month-old funnel, not Cherry Tree traffic quality. |
The Honest Summary
Cherry Tree delivered AI citations in the first week — on a domain that was only 6 months old. Within 8 weeks, 20–24% of all confirmed traffic came from AI conversations. Organic Google search grew 527% over the same period — same content, no separate SEO strategy, no backlinks, no ad spend.
Four AI platforms are now actively citing the content. 65 of 269 articles have confirmed citations. Factor in dark AI traffic that GA4 misses entirely, and the real AI share is closer to 40%.
The discovery engine works. The conversion funnel — built after this measurement window closed — is a separate problem, and not a Cherry Tree one. Seresa now has traffic to convert! This complete and detailed analysis has only been possible with the help of Claude AI and the data collected first party via our Transmute Engine service.
Chart 1 of 8 — AEO Traffic Over Time
AEO traffic appeared before Google finished crawling.
Five confirmed AEO sessions from three AI platforms in the first partial week — on a domain that was 12 days old. Traditional SEO: 6–12 months. Cherry Tree: 4 days.
40%
Combined AI traffic confirmed + dark
101
Confirmed AEO sessions
527%
Organic growth wk1 → wk9
12 days
Domain age at first AI citation
Confirmed AEO (blue) vs Dark AI estimate (orange) · 9 weeks · seresa.io · Transmute Engine data · Partial week one. Domain registered 11 March 2026.
Google hadn’t finished crawling. Claude, ChatGPT, and Perplexity were already citing.
The Dark AI Discovery
For every confirmed AEO session, 1.5 more are hiding in “direct”
Traditional analytics classifies AI-referred traffic as “direct” whenever an AI app strips referrer headers — which most mobile AI apps now do. The Transmute Engine preserves the raw referrer alongside the resolved source, which lets us cross-reference both and detect the AI traffic that GA4 would simply lose.
We define a dark AI session by four simultaneous signals: first-time visitor, no referrer, no source attribution, and a landing page in /blog/ or /seed/. These four together produce a session that behaves exactly like a confirmed AEO visitor — long engagement, deep landing pages, no commercial intent — but with no breadcrumb back to the AI that sent them.
The Numbers
Most of the AI traffic never identifies itself.
101
Confirmed AEO · 15.4%Sessions with a verified AI referrer — claude.ai, chatgpt.com, perplexity.ai, doubao.com. Captured by the Transmute Engine’s raw-referrer field.
+156
Dark AI estimateFirst-visit sessions with no referrer landing on deep blog/seed pages. The engagement fingerprint matches confirmed AEO — same content, same behaviour, no breadcrumb.
257
Combined · 39.3%Total estimated AI-driven sessions across 56 days. Roughly two in every five visits to seresa.io now originates from an AI conversation.
2.5×
The multiplierReal AI reach over what confirmed-only attribution would show. A GA4-only analyst would have reported 15.4%. The truth is closer to 40%.
Chart 2 of 8 — The Traffic Didn’t Drop, It Went Dark
The confirmed AEO “collapse” in May is a measurement problem, not a citation problem.
Confirmed AEO fell from 17/week to 3/week in May. Dark AI rose by the same amount. Combined AI share held steady at 37–42%.
39.3%
True AI traffic share captured by Transmute Engine raw-referrer
15.4%
What GA4 reports confirmed referrers only — misses the dark half
Without server-side raw-referrer capture, this entire channel is invisible. Every AEO case study using GA4 alone is understating AI traffic by at least half.
2.5× — the dark AI multiplier
For every confirmed AEO session, there are approximately 1.5 more sessions hiding in the “direct” bucket. Without server-side raw-referrer capture, this entire channel is invisible. Every AEO case study that doesn’t investigate the direct/no-referrer cohort is understating AI’s contribution by at least half.
Chart 3 of 8 — Platform Analysis
Four platforms. Three active. One dominant.
Claude leads at 53% — but each platform cites different content. Claude: technical comparisons. ChatGPT: narrative problem/solution. Perplexity: direct-answer seeds. Same library, three different entry points.
4
Platforms citing Claude, ChatGPT, Perplexity, Doubao
65 / 269
Articles cited 24.2% of the library
17
Peak week unique pages cited
1
Non-Western platform Doubao — APAC signal
Claude
54 sessions · 53%Dominant. Spiked to 16/week (Apr 20). Prefers structured technical comparisons that definitively answer “which option should I choose.”
ChatGPT
34 sessions · 33%Steady 4–6/week. Appends utm_source=chatgpt.com. Favours narrative and trend content with a clear problem/solution arc.
Perplexity
13 sessions · 13%Sporadic but real. Peaked at 6/week (Apr 27). Loves direct-answer seeds and product comparison pages.
Doubao (ByteDance)
1 session · the early signalFirst non-Western AI citation. One session today — but ByteDance serves 1.7B users across Asia. This is the early signal.
Chart 4 of 8 — Geographic Reach
Ten countries. AI cites across borders.
AEO sessions reached visitors in 10 countries within 8 weeks, with no paid distribution and no SEO authority. The 84% “Unknown” rate isn’t a Transmute Engine gap — it’s AI platforms stripping geo data upstream.
AEO session geography · 84% country unknown (AI strips geo-IP data) · 10 confirmed countries
Chart 5 of 8 — The “Other” Decomposition
“Other” isn’t a junk drawer. It’s where the real story hides.
451 sessions were classified as “Other” — 68.9% of all traffic. We broke them apart. The single largest sub-channel hiding inside? Dark AI. Bigger than confirmed AEO. Bigger than organic search.
451
“Other” sessions 68.9% of all traffic
156
Suspected dark AI larger than confirmed AEO (101) and organic (116)
78.9s
Dark AI engagement 51% longer than confirmed AEO
“Direct”
GA4 classification invisible without server-side data
The largest single channel inside “Other” is AI traffic that stripped its own referrer. GA4 calls it “direct.” Transmute Engine calls it what it is.
Engagement Quality
AEO visitors verify. Organic visitors explore.
AEO engagement (52.1s, 30.6% scroll, 62.4% engaged) is respectable in absolute terms — but organic search beats it on every single metric. That’s the nature of the channel, not a content quality failure. AEO visitors click through from an AI’s citation, confirm the source is legitimate, and return to their AI conversation. Organic visitors arrive without that pre-summary, so they read.
Both channels matter. AEO is the discovery loop; organic is the deep-read loop. The same Cherry Tree content drives both — that’s the structural win.
| Metric | AEO | Organic | Dark AI |
|---|---|---|---|
| Avg engagement time | 52.1s | 64.8s ↑ | 78.9s |
| Avg scroll depth | 30.6% | 49.5% ↑ | 32.5% |
| Engagement rate | 62.4% | 80.5% ↑ | — |
| Multi-page rate | 13.9% | 15.5% ↑ | — |
Chart 6 of 8 — Platform Behaviour
Each AI platform cites different content. That’s the structural advantage.
- ClaudeTechnical comparisons. “Which should I choose?” content. 53% of all AI sessions.
- ChatGPTNarrative problem/solution. E-commerce how-tos. Steady 4–6/week.
- PerplexityDirect-answer seeds. Product comparisons. Sporadic but high-intent.
- Doubao (ByteDance)First non-Western citation. One session today. 1.7B users tomorrow.
Four AI platforms. Four different content appetites. One Cherry Tree library that feeds all of them.
Chart 7 of 8 — AEO vs Organic Growth
The same content drives two compounding channels.
Cherry Tree articles climbed from 2 to 23 organic sessions/week over the same 8 weeks — same content, no separate SEO strategy, no backlink building.
2
Organic wk1 sessions / week
23
Organic wk9 sessions / week
527%
Growth no SEO strategy, no backlinks
1 → 2
Content pipeline → channels one library, two discovery engines
The article Claude cites today is the article Google ranks tomorrow. Same page. Two discovery engines.
527%
Organic search growth, week 1 → week 9. One content pipeline. Two compounding channels. No SEO budget.
Chart 8 of 8 — The Organic Story
The same content that gets cited by AI gets ranked by Google.
The article that ranks for a long-tail Google query is the same article Claude cites in conversation. One pipeline. Two channels. Compounding returns.
6.3 on a 10-week-old domain with zero brand, zero backlinks, and no conversion funnel yet built. The scores Cherry Tree controls — citation speed, citation breadth, measurement capability, platform diversity — average 8.3 / 10.
Independent assessment. First-party data. No vanity metrics. The full methodology and query logs are available on request.
The Verdict
Cherry Tree solves discovery. In 4 days, not 6 months. The conversion funnel is a separate build.
The data unambiguously confirms that Cherry Tree produces structurally AI-citable content. 24% of the library is being actively cited. Three major platforms are participating. A fourth (Doubao) is emerging. Organic search is compounding on the same content. The discovery engine works. On a domain that didn’t exist 10 weeks ago.
The conversion funnel was built after the measurement window closed. The 4 / 10 conversion score reflects a funnel that wasn’t there yet — not traffic that failed to convert.
Recommendation for a Cherry Tree Client
Who should buy it — and who shouldn’t
Buy it if…
- You want AI citations within days, not months — on any domain age
- Your product benefits from being cited as an authority in AI conversations
- You’re willing to build the conversion funnel (or with Seresa’s help) once discovery traffic establishes
- You want one content investment that compounds in both AI citations and Google organic — same pages, two channels
Don’t buy it if…
- You need last-click-attributed leads this quarter
- You don’t have (or won’t build) a conversion funnel downstream
Addendum — data quality verdict: 8.5 / 10
The Transmute Engine data is substantially better than what a GA4-only setup would provide for this type of analysis. Raw referrer preservation, server-side collection, and event-level engagement metrics are the three features that made the dark AI analysis possible. A GA4-only analyst would have concluded “AEO traffic dropped in May” and stopped there — because the data to see otherwise wouldn’t exist. For any client considering the Seresa stack: the data infrastructure is arguably more impressive than the content pipeline at this stage.
Get cited by AI — starting in week one.
Cherry Tree builds content that AI engines cite — fast. The data from this case study was generated on a brand-new domain. The same pipeline works for any technical or B2B subject where AI conversations are how prospects already research. Get AEO running for you today for less than a coffee a day.
