.svg){kind=logo}
Beedeez is a French LMS built for field teams. 2 million users, 55 countries, and customers like Leroy Merlin, Würth and the French Navy. Solid brand on paper. Near-invisible in answers from ChatGPT, Perplexity and Gemini back in November 2025.
The challenge: visible by name, invisible by intent
Search for Beedeez by name, and the brand showed up 100% of the time across every AI engine we tested. Ask something generic like "best LMS to train frontline teams", and visibility crashed to 12%.
In other words: Beedeez was known, but not recommended.
Three reasons:
- AI models were pigeonholing the brand as mobile learning, not as an LMS for deskless workers.
- On the prompts that matter to a buyer, 360Learning and Docebo were showing up instead.
- Share-of-voice gap with the leader sat at 1 to 1.75.
The key insight: a brand that ranks on traditional search can still be completely absent from AI answers. These are two separate battles, with different signals.
Our diagnosis
The audit surfaced three compounding blockers.
- Broken technical foundations. 600+ pages with duplicate H1s, unresolved 301 and 404 redirects, images missing ALT text. Zero schema markup. No llms.txt file. No working EN version. AI crawlers and Google were both struggling to parse the site.
- Muddled semantic positioning. Existing content kept using "mobile learning", "microlearning", "training app". Outdated terms that no longer matched the product. LLMs were drawing the wrong conclusions.
- Dormant editorial engine. Few recent articles, no finished industry pages, no high-intent formats like "Alternatives to [competitor]". AI engines had nothing fresh to pull from.
A 6-month engagement, in three phases
Six months structured around three moves: lay the foundations, publish hard, consolidate. Each phase had one priority lever and measurable deliverables.
Phase 1 (M1 to M2): Technical foundations. 600+ pages cleaned up, full schema markup rollout (FAQPage, Article, SoftwareApplication, AggregateRating), llms.txt created, EN version launched. Result: AI visibility moved from 12% to 17%. Modest, but the site is finally readable.
Phase 2 (M3 to M4): Editorial unlock. A prioritization matrix was built from BOFU prompt analysis. First wave of aggressive publishing: 7 new articles, 2 refreshed case studies, 3 rebuilt industry pages. Gemini added to tracking. Result: +72% visibility in 30 days. That's the turning point.
Phase 3 (M5 to M6): Consolidation. 8 industry pages shipped, 7 flagship case studies updated, author pages with Person schema for stronger E-E-A-T. Semantic repositioning finally locks in: AI engines now tie Beedeez to "LMS for field teams" instead of "mobile learning".
"In 6 months, Vydera structured our visibility on Google and across AI engines. We went from challenger to #2 in the French LMS market."
Hamza Sbaa, CMO, Beedeez
The competitive trajectory
In 5 months, Beedeez caught up with and overtook Docebo, TalentLMS, iSpring and Moodle. The gap to 360Learning narrowed from 14.8 points to 10.5. This is what a challenger climb looks like.
Results after 6 months
Across the 30 strategic prompts we tracked continuously on ChatGPT, Perplexity and Gemini:
- AI visibility multiplied by 4 (from 12% to 49%)
- Moved from 4th to 2nd in the GEO ranking
- Citation rate doubled (from 26% to 53%)
- Monthly AI traffic up 56% (from 84 to 131 visits)
- AI sentiment steady at 4.3 out of 5 throughout the period
The most important number sits inside the prompt typology breakdown. On BOFU prompts (purchase intent), visibility went from 31% to 52%. That's where buyer shortlists get locked in. On informational prompts, visibility moved from 26% to 48%.
The MOFU segment (comparison stage) is still open at 28%, dominated by competitors who have shipped heavy comparison content. That's the priority for the next 6 months.
Deliverables: 40+ briefs produced, 15 pieces of content published, 8 industry pages shipped, 600+ pages audited and technically fixed, 30 prompts tracked continuously, 100% schema markup coverage in EN and FR.
3 takeaways
1. Brand recognition won't save you with AI engines. A brand can have 100% recall when searched by name and still be completely absent from generated answers on use-case queries. Two different battles.
2. Technical foundations make the rest possible. Without the cleanup in months 1 and 2, none of the phase 2 content would have delivered the same results. The acceleration in M3-M4 only happens because the site is finally readable.
3. Publishing cadence is the number one lever. The correlation between shipping articles and AI visibility is direct, fast, and reproducible. GEO isn't a strategy problem. It's a cadence problem.
Want to show up in AI answers when buyers come looking? Let's talk.
What is GEO (Generative Engine Optimization)?
GEO is the discipline of getting your brand surfaced in answers generated by AI engines: ChatGPT, Perplexity, Gemini, Claude. While classic SEO aims to rank in Google's result pages, GEO aims to be cited as a source or recommended when a user asks a question. Both share common foundations (content, schema markup, authority) but optimize for different signals.
How long does it take to see results from GEO?
Early signals appear within 30 to 60 days if technical foundations are clean. With Beedeez, the first measurable inflection hit in month 4 with +72% AI visibility in 30 days, after 2 months of technical cleanup and 1 month of content production. Expect 6 months for a consolidated trajectory across most strategic prompts. Any GEO provider promising results in 2 weeks without existing technical foundations is lying.
What's the difference between SEO and GEO?
SEO optimizes for deterministic ranking algorithms (Google, Bing) that return a list of 10 blue links. GEO optimizes for probabilistic LLMs that generate a single answer by synthesizing multiple sources. The signals that matter are different: SEO values backlinks and exact rankings, GEO values citability (structured content, FAQs, hard data, schema markup) and semantic consistency across the web. Strong SEO content doesn't guarantee strong GEO content. Both strategies need to run in parallel.
How do I know if my brand is visible on ChatGPT and Perplexity?
You need to track a panel of strategic prompts over several weeks, either manually or through a dedicated tool like Meteoria, Profound, or AI Monitor. Measure three core indicators:
- Appearance rate - how many prompts surface your brand
- Citation rate - how many cite your domain as a source
- Sentiment - positive, neutral, or negative framing
Without measurement, there is no way to know whether a GEO action is actually moving the needle.
Does GEO replace SEO?
No. GEO complements SEO, it doesn't replace it. Google remains the top organic acquisition channel for most B2B brands, and Google AI Overviews pulls directly from pages that rank well in classic SEO. A site that ignores SEO will struggle to feed AI engines, which lean heavily on Google results to build their answers. The right approach is SEO + GEO in parallel, with distinct teams and KPIs but aligned strategies.
.jpeg)
.jpg)
.jpeg)
