Entry #13 · May 5, 2026
Three weeks ago we ended an article with a question. If 95% of AI-mediated conversations about Genymotion are about free usage and mobile gaming, can we shift that ratio toward enterprise? We made changes to the AI site and we just looked at the data.
The traffic on the chatbot on the website hasn’t shifted. On the AI site however, it has.
Buyer/professional content reads went from 35.9% to 45.0% (+9.1pp). The chatbot stayed flat at 7–8% buyer rate.
Two channels, two audiences.
Where we left off
In article #11 we shared traffic numbers. On the Genymotion website, the on-site chatbot recorded 667 conversations in March. Fewer than 20 of them showed clear commercial intent, under 5%.
The rest were people trying to run TikTok on their PC. The rest were people trying to set their device language to Chinese. The rest were people trying to play Minecraft on Steam.
The marketing team targeted CI/CD teams. The marketing team targeted mobile security testers. The marketing team targeted cloud-deployment buyers.
Those audiences were less than 5% of the conversations the chatbot recorded.
That same article also showed a different pattern on the AI side. A pricing evaluation from Madrid happened with eleven ChatGPT-User fetches across five rounds. Those fetches ended on the full pricing breakdown.
A late-night deliberation from the US happened. macOS compatibility checks happened from multiple continents.
These were buyer-pattern sessions. All sessions happened in ChatGPT. The website never recorded them.
We asked at the end of that article whether the AI site could push the ratio. Could we get AI platforms to recommend Genymotion to buyers, not just describe it to hobbyists?
We can now answer that with three more weeks of data. We compared a 14-day window in mid-March (the article #11 baseline) with a 14-day window from April 21 to May 4.
Both for the chatbot on genymotion.com and for the AI site at rozz.genymotion.com.
Chatbot rate: flat
The Rozz chatbot now classifies every conversation with a buying_intent flag (yes/no). The definition is broader than the manual “clear commercial intent” we used in March. The absolute numbers are higher.
Applied consistently before and after, the classification gives a consistent measure.
| 14-day window | Conversations | Classified | buying_intent='yes' | Rate |
|---|---:|---:|---:|---:|
| Mar 11 – 24 | 338 | 335 | 24 | 7.2% |
| Apr 21 – May 4 | 273 | 210 | 17 | 8.1% |
Plus or minus a percentage point. Classifier coverage was partial in the late-April window because the pipeline is still rolling out.
It applied uniformly to whatever it classified. That uniform application makes the rate comparable.
The broader monthly picture included February 7.9%. The broader monthly picture included March 7.6%. The broader monthly picture included April 10.3% (with classifier coverage still partial).
The chatbot’s buyer rate has stayed between 7% and 10% for months. Same hobbyist questions appear about Minecraft. Same hobbyist questions appear about TikTok. Same hobbyist questions appear about Chinese language packs.
The website’s audience composition is essentially what it was when we wrote article #11.
If the goal was to convert the chatbot’s hobbyist audience into a buyer audience, that didn’t happen. The website is still mostly hobbyists.
The AI site shifted toward enterprise
We classified every page that ChatGPT-User fetched in both windows by what the page actually contains.
Pricing pages count as “buyer / professional”. Free-tier pages count as “buyer / professional”. License pages count as “buyer / professional”.
Install guides count as “free user”. System requirements count as “free user”. “I can’t find the Play Store” content counts as “free user”.
We set aside discovery pages because they are navigation, not content. Discovery pages include the homepage and topic listings.
| Audience cluster (content reads on the AI site) | Mar 11–24 | Apr 21–May 4 | Δ | |---|---:|---:|---:| | Buyer / professional (pricing, cloud, CI-CD, security testing) | 264 (35.9%) | 385 (45.0%) | +9.1pp | | Free-user (install help, troubleshooting, requirements, compatibility) | 294 (39.9%) | 270 (31.5%) | −8.4pp |
Six weeks ago, free-user content was the bigger bucket. Today, buyer/professional content is.
By design, as we added enterprise content between the two windows, some of the shift is supply-driven.
The new security-testing page (use-burp-suite-with-genymotion-desktop) got 89 reads in 14 days. In March, that page didn’t exist.
Cloud product pages grew from 48 to 71 reads. CI/CD content went from 5 to 12.
If a site publishes more enterprise pages, AI bots will fetch more enterprise pages. This shift is supply-driven.
What stayed flat is not supply-driven. Reads of what-pricing-plans-are-available-for-genymotion were 42 in March. Reads of what-pricing-plans-are-available-for-genymotion were 47 in late April.
The existing buyer pages aren’t being read dramatically more.
The likely explanation is that ChatGPT-User didn’t fetch content on the Genymotion AI site when enterprise queries were made before the corresponding pages existed. Once the corresponding pages existed, AI bots fetched them.
On May 1–2, three separate sessions paired the new Burp Suite page with i-recently-upgraded-virtualbox-and-genymotion-no-longer-work. Mobile pen-testers were hitting a real bug.
What we changed on the AI site
In the six weeks between the two windows, we shipped changes to how the AI site ranks and presents Q&As.
Our numbers indicate that ChatGPT-User fetches the homepage of rozz.genymotion.com for roughly 25% of its retrievals. We chose to change two things that AI platforms are known to read.
AI platforms read what is at the top of the page. AI platforms read what is inside the FAQPage JSON-LD.
Three weeks ago, the homepage led with “is genymotion free?”. The FAQ was ranked by raw retrieval count from CloudFront logs.
Whichever Q&As AI bots fetched most often got promoted. Hobbyist queries dominated. Hobbyist Q&As led the homepage.
When AI platforms read the homepage to answer enterprise queries, AI platforms found mostly hobbyist content.
We made three sets of changes.
Buying-intent ranking
The chatbot already classifies every conversation by buying_intent. We applied that signal to the AI site.
If a Q&A’s origin conversation is tagged buyer, the Q&A gets a ranking boost in the FAQ selector.
We started at +0.25 on Apr 17. The first round didn’t shift the rankings enough. We raised it to +0.5 the same day.
On Apr 20 we unified the HTML FAQ and the FAQPage JSON-LD to use the same boosted selector. Before this unification, they were picked independently.
The JSON-LD just took the first 10 Q&As in database order. That selection was wrong.
The JSON-LD that AI bots read directly was the least curated part of the site.
A “For Enterprise Buyers” section, then a reversal
On Apr 17 we also added a dedicated “For Enterprise Buyers” section to the homepage and llms.txt.
Five days later we audited what it actually contained. Only ~27% of the Q&As in that section were actually about enterprise topics.
The rest were consumer content. The buying_intent classifier had over-tagged this consumer content. This over-tagged content included install/uninstall Linux. This over-tagged content included “light use not gaming”. This over-tagged content included Bluestacks comparisons.
We removed the section on Apr 22.
Labeling enterprise content has a precision problem when the upstream signal is noisy. Ranking the same content higher in the shared FAQ doesn’t have that risk.
A borderline Q&A that gets boosted is still inside a generic FAQ. The Q&A is not under a heading that overstates its category.
Editorial overrides
On Apr 21 we shipped two manual controls.
One control pins specific Q&As to the top of the FAQ. The other control promotes specific topics to the front of the topic directory.
The Genymotion homepage now leads with Network & Security Config. The homepage now leads with Mobile Test Automation. The homepage now leads with CI/CD Automation. The homepage now leads with Cloud Deployment Options.
That ordering is a curated decision. That ordering is not an algorithmic decision. The algorithm picks the long tail. Editors pick the front.
Two structural cleanups also mattered.
Each Q&A used to appear in 4–7 topic cards because of keyword-based inheritance. 86% of Q&As were duplicated across multiple topics.
A new classifier (Apr 22) put each Q&A in 1–2 canonical buckets. Duplication is now at 2%.
On Apr 23 the topic taxonomy became persistent across crawls. URLs stopped changing on every regeneration. This is the URL-churn problem we wrote about in article #12.
Claude Code activity grew
In article #9 we documented the first ever Claude-User session on the AI site.
That session had 14 requests over six days in late March. The session was mostly Claude Code. We treated the session as an early signal.
The signal continued. In the past two weeks we logged 26 Claude-User hits.
24 hits came from Claude Code. Two sessions are notable:
| May 1, 19:49 UTC | May 2, 10:28 UTC | |---|---| | 6 fetches in 46 seconds | 19 fetches in 70 seconds | | Index → Cloud Deployment Options → QnA index → what-pricing-plans-are-available | Index → Cloud Deployment Options → Virtual Device Management → Android Dev Integration → costs-of-cloud-and-billing → can-i-run-my-apk → arm-support-saas → credit-card-trial → simultaneous-devices → saas-vs-desktop → how-can-i-run-locally → cloud-marketplace-pricing → gpu-arm-support → bluetooth → desktop-requirements |
The second session matches a procurement evaluation pattern. Cloud deployment and billing model appear. ARM-on-SaaS appears. Trial requirements appear. SaaS-versus-Desktop economics appear. Marketplace pricing appears.
Someone or something traversed the site in the order a procurement evaluator would use. That sequence covered cost. That sequence covered scaling. That sequence covered technical compatibility. Then the sequence returned to cost from a different angle.
In the entire 14-day window, Claude-User fetched zero install-help pages. Claude-User fetched zero troubleshooting pages. Claude-User fetched zero “can’t find” content.
Of 11 actual content reads (excluding navigation), 50% were enterprise-evaluation pages. Of 11 actual content reads (excluding navigation), 33% were cloud-product pages.
The sample is small (n=11). It is consistent with every other measurement.
So, did the buyer ratio shift?
It depends on the channel. The AI site shifted. The website didn’t. The two channels produced different results.
The genymotion.com website still draws its old audience. The audience includes hobbyists. The audience includes gamers. The audience includes language-pack tweakers.
The website has the same 7–8% buyer rate as before. Nothing the AI site did changed who arrives at genymotion.com.
On the AI side, different sessions are now appearing. A Claude Code procurement session appears. A mobile pen-tester pairing Burp Suite with VirtualBox troubleshooting appears.
They asked an AI a question. The AI fetched the AI site.
> This is what’s new. The AI site is read by an audience that doesn’t show up in numbers on the website, especially if they’re dwarfed by the numbers of the free users. However, the marketing team’s enterprise targets (CI/CD, mobile security, cloud at scale) do appear in the AI site’s logs.
Buyers query AI platforms. AI platforms read the AI site to answer them.
What we cannot yet claim
We see what AI platforms read. We don’t see what AI platforms say.
Did ChatGPT and Claude actually recommend Genymotion in their answers during those reconstructed sessions? Did the buyer who asked “what cloud emulators integrate with our CI pipeline” get “Genymotion” as the answer?
Those answers are separate measurements. The citation tracker handles them. We’ll come back to it later.
A sample-size note applies. Claude-User content reads are small. Claude-User content reads are a representative observation, not a statistical claim.
The ChatGPT-User numbers are much higher. The ChatGPT-User numbers include 856 content reads in 14 days. The ChatGPT-User numbers show the same pattern.
Another note applies. We can’t say from this data that buyers stopped visiting the website.
We can say that the buyer-pattern sessions we see in AI logs don’t appear in the chatbot logs. Would those buyers have visited the website in a pre-AI world? Is AI now substituting for that visit?
Cross-deployment data is required to answer with confidence.
Get this for your company
Rozz gives visibility into the AI conversations happening about your product. Rozz gives tools to influence what AI recommends.
Data source and period
Data source includes on-site chatbot logs for genymotion.com (buying_intent classifier). Data source includes CloudFront access logs for rozz.genymotion.com.
Two 14-day windows exist. The baseline window is Mar 11–24, 2026. The post-changes window is Apr 21–May 4, 2026.
ChatGPT-User and Claude-User reads are classified by content audience cluster.
Author
Author: Adrien Schmidt, CEO, ROZZ
Serial tech entrepreneur with 10+ years experience building AI systems including Aristotle (conversational AI analytics) and products for eBay and Cartier. Previously founded Squid Solutions and built AI products like Aristotle, the conversational big data analytics chatbot, and an AR jewelry try-on device for Cartier.
Metadata block
May 5, 2026 | Data period: Mar 11–24 vs. Apr 21–May 4, 2026 (two 14-day windows)