Mastering AI Search: A Case Study on Optimizing E-commerce Product Pages for ChatGPT, Gemini, and Perplexity
A practical, measured guide to generative engine optimization.
Estimated Read Time: 20 minutes
Last updated: February 2026
Many e-commerce businesses dedicate significant resources to SEO, ensuring their product listings shine on Google. Yet, a seismic shift is underway: a growing number of consumers are bypassing traditional search engines, turning directly to AI conversational agents like ChatGPT, Gemini, and Perplexity to ask for product recommendations.
If your product pages aren't visible to these AI-powered assistants, you're missing out on a critical new sales channel.
This was precisely the challenge my e-commerce client faced. In this in-depth article, I'll walk you through the complete methodology I employed to make a PrestaShop product page discoverable and cited by leading AI engines. You'll see the exact steps taken, alongside tangible before-and-after performance metrics.
The Silent Problem: Product Pages Invisible to AI
My client, a PrestaShop retailer of trail running shoes, had a flagship product – a technical terrain trail shoe – performing well in Google search results:
- Ranked 8th for "technical terrain trail shoe"
- Generating roughly 120 organic visits each month
- Boasting a healthy 3.2% conversion rate
On the surface, everything appeared satisfactory. However, I felt compelled to investigate a new frontier.
The AI Visibility Audit: A Wake-Up Call
I posed a straightforward question to three prominent conversational AIs:
“What is the best trail running shoe for technical terrain in 2025?”
The outcome was stark:
| AI Engine | Product mentioned | Brand mentioned | Link to page |
|---|---|---|---|
| ChatGPT (GPT-4o) | No | No | No |
| Gemini | No | No | No |
| Perplexity | No | No | No |
Not a single mention.
Instead, all AIs consistently recommended well-established brands like Salomon, Hoka, and La Sportiva. Despite its strong Google presence and clear relevance, my client's product page was utterly absent from the AI search landscape.
This eye-opening discovery marked the genesis of our Generative Engine Optimization (GEO) initiative.
Understanding GEO: The New Paradigm for E-commerce
GEO is the strategic process of optimizing your digital content to be accurately interpreted, chosen, and referenced by AI answer engines, extending beyond simple indexing by conventional search platforms.
The core distinction is crucial:
- Traditional SEO focuses on algorithms that present a list of blue links.
- GEO targets language models that synthesize a single, authoritative answer from diverse sources.
With SEO, ranking #8 might still bring traffic. In the realm of GEO, you either feature in the AI's direct answer or you effectively don't exist. There's no "second page" in generative AI.
Why Conventional Product Pages Fall Short in GEO
Most e-commerce product pages are designed to convert visitors who have already arrived. They typically include:
- A title honed for keywords
- Feature bullet points
- Product imagery
- An "Add to cart" button
However, they often lack the crucial elements an AI needs to confidently recommend a product:
- Verifiable claims with explicit sources
- Detailed context for usage
- Transparent comparisons with competitor products
- Structured data that AIs can easily parse
- Credibility signals (e.g., in-depth reviews, independent tests, certifications)
The Complete GEO Playbook: 7 Steps to Transform Your Product Page
Here's a step-by-step account of the precise actions I took.
Step 1 — Initial AI Visibility Assessment
Before making any changes, I executed a thorough GEO visibility audit. Beyond the generic query mentioned earlier, I compiled 15 diverse conversational prompts, each representing a distinct user intent:
Informational Queries:
- “What trail shoe works best for rocky terrain?”
- “Compare technical trail shoes 2025”
- “Which trail brand offers the best for rugged terrain?”
Transactional Queries:
- “Buy affordable technical terrain trail shoe”
- “Best trail shoe for value”
Specific Queries:
- “Trail shoe with Vibram sole for ultra-trail running”
- “Review of trail running shoes for rocky terrain”
For each query, I meticulously recorded in a spreadsheet:
- The full response from each AI
- Whether my client's product or brand was cited
- Which competing products or brands were mentioned
- The original sources cited by the AI (particularly visible with Perplexity)
Audit Outcome: Across 45 query combinations (15 queries × 3 AIs), my client's product registered zero appearances. Competitors frequently cited included Salomon (38 mentions), Hoka (31 mentions), La Sportiva (27 mentions), and Scarpa (19 mentions).
Key takeaway: AI models tend to favor recommendations based on what they are most familiar with – meaning brands and products heavily represented in their training datasets and real-time information sources.
Step 2 — Dissecting Competitor Success in AI Search
To strategically position my client's product within AI recommendations, I first needed to understand why competitors were succeeding. I analyzed the product pages of models frequently cited by the AIs, uncovering common threads.
Recurring characteristics of successful competitor pages:
- Rich, editorial-style content: More than just specifications, these pages featured detailed narratives explaining the ideal user, optimal conditions, and distinct advantages.
- Comprehensive structured data: Extensive Schema.org Product markup, including
aggregateRating,review,offers,brand, and evenisSimilarToproperties. - High-authority third-party mentions: Reviews in specialist publications (e.g., RunRepeat, Outdoor Gear Lab), comparisons on expert blogs, and discussions within active community forums.
- Structured and insightful customer reviews: Not just brief "5-star" ratings, but testimonials detailing specific use cases and experiences.
- Robust semantic interlinking: The product page was part of a broader content ecosystem, linked to relevant buying guides, blog posts, and FAQs, which collectively bolstered the site's topical authority.
Deficiencies in my client's original page:
- A mere 80-word product description, largely confined to technical specifications.
- Minimal structured data, limited to PrestaShop's default settings.
- No external mentions from authoritative third-party sites.
- Only four short customer reviews.
- A complete absence of supporting editorial content on the blog.
Step 3 — Overhauling the Product Page for AI (and Human) Engagement
This stage proved pivotal. I undertook a complete overhaul of the product page, integrating core GEO principles.
Before: The Initial Product Page
TrailForce X1 — Technical terrain trail shoe
Vibram MegaGrip sole
6mm drop
Weight: 310g
Breathable mesh upper
Rock guard protection
Colors: black/red, grey/blue
Ideal for technical trails and rugged terrain.
A concise 80 words. Devoid of context, comparisons, or verifiable claims.
After: The GEO-Optimized Product Page
Below is the refined page structure. Each section is detailed, with an explanation of its strategic purpose.
1. A Semantically Richer Title
TrailForce X1 — Elite Trail Shoe for Rocky & Rugged Terrain | Expert Review & 2025 Ratings
Why: AI engines prioritize titles as strong relevance indicators. Incorporating phrases like "rocky & rugged terrain" and "Expert Review & 2025 Ratings" significantly improves the likelihood of matching conversational queries.
2. An "Expert Insight" Introduction
The TrailForce X1 is a high-performance trail running shoe engineered for the most demanding technical terrain, including rocky paths, scree fields, and challenging singletracks. Featuring a robust Vibram MegaGrip sole and a 6mm drop, it masterfully combines superior grip, essential protection, and agile ground feel—qualities often found in models costing 30% to 50% more. Our rigorous testing across 200 km of varied Alpine trails highlighted its exceptional downhill stability on technical ground and remarkable durability, with the sole remaining in pristine condition after extensive use on abrasive surfaces.
Why: This introductory paragraph is packed with specific, credible assertions ("200 km tested," "30% to 50% cheaper"), provides clear use-case context ("rocky paths, scree fields"), and offers an implicit comparative assessment. AIs favor this kind of content because it directly addresses user questions with factual, detailed information.
3. A "Who Is This Shoe For?" Segment
Ideal User: Intermediate to advanced trail runners tackling technical routes (scree, roots, rocks) across distances from 15 km to 80 km.
Shines When: You seek maximum traction on wet rock, impact protection without sacrificing responsiveness, and exceptional value within a controlled budget.
Not Suited For: Flat forest trails (its rigidity is overkill), ultra-trails exceeding 100 km (cushioning may be insufficient for very long distances or runners over 75 kg), or runners preferring a higher heel drop.
Why: This section is paramount for GEO. When users query AIs about "which shoe for X scenario," the AI seeks content that clearly delineates appropriate use cases. Openly stating the product's limitations also acts as a powerful credibility signal, a trait AIs are increasingly programmed to value.
4. Technical Specifications, Reimagined with Context
Rather than a bare list of specs, I transformed each feature into a narrative explaining its tangible benefit on the trail:
| Feature | Detail | Impact on the Trail |
|---|---|---|
| Sole | Vibram MegaGrip, 5mm multi-directional lugs | Unrivaled traction on wet rock – often compared to Salomon Speedcross on dry, but significantly outperforms it on slick surfaces. |
| Drop | 6 mm | A versatile balance, promoting uphill power and comfortable, sustained downhill performance. |
| Weight | 310 g (size 42/US 9) | Standard for technical footwear – slightly heavier than a Hoka Speedgoat, but justified by superior rock guard protection. |
| Protection | Full rock guard + anti-perforation plate | Empowers confident traversal of scree fields and jagged terrain where lighter shoes necessitate extreme caution. |
| Upper | Technical breathable mesh with welded reinforcements | Offers ample breathability in warmer conditions, while reinforcements bolster durability against trail abrasions. |
Why: Explicit comparisons with well-known products (Salomon, Hoka) are a potent GEO accelerator. By relating the product to established references, the AI can effortlessly integrate it into its existing knowledge framework and articulate its unique position.
5. A Strategic FAQ Section
Is the TrailForce X1 suitable for ultra-trail events?
It's an excellent choice for technical terrain up to 80 km. For distances beyond this, particularly for runners over 75 kg, the cushioning might prove insufficient. For extreme ultra-distances, we recommend [link to another model in the range].How does it compare to the Salomon Speedcross 6?
The Speedcross 6 excels on soft, loose terrain like mud, but offers less versatility on rocky surfaces. The TrailForce X1 delivers superior stability and protection on rocks, typically at a price point approximately €40 lower.What size should I order?
The TrailForce X1 generally fits true to size. For trail running, it's common practice to size up by half a size to prevent toe discomfort on descents, especially if you're between sizes.
Why: FAQs are among the most effective formats for GEO. The questions directly mirror typical conversational AI queries. AIs effortlessly identify these sections and frequently extract them as direct sources for their answers.
6. A "Tester Endorsements" Block with Direct Quotes
4.6/5 — Consolidated rating across 47 verified purchases
"I conquered the Trail des Aiguilles Rouges (58 km, 3,200 m elevation) in these shoes. The grip on wet granite slabs was flawless. The best discovery of my season!" — Julien M., Ultra-Trail Runner, Chamonix
"I transitioned from Hoka Speedgoat 5. While a bit heavier and less cushioned, the grip and protection on rocky terrain are in a different league." — Sophie L., Intermediate Trail Runner, Grenoble
"Unbeatable value for technical trails. My only minor critique: the lacing system could be refined." — Marc D., Field Tester for TrailSession.fr
Why: Detailed reviews, especially those with contextual information (event names, comparisons to other models, or even a minor critique), serve as strong credibility signals for AIs. Reviews that explicitly compare the product to a known competitor are particularly valuable for helping the AI contextualize and differentiate the offering.
Outcome: The product page evolved from a meager 80 words to a substantial 900 words of valuable, structured content, purpose-built for conversational queries.
Step 4 — Implementing Advanced Structured Data
PrestaShop’s native structured data is usually basic and insufficient for the demands of GEO. I implemented a significantly enriched Schema.org markup:
{
"@context": "https://schema.org",
"@type": "Product",
"name": "TrailForce X1 — Technical terrain trail shoe",
"description": "Trail shoe designed for rocky and rugged terrain, equipped with a Vibram MegaGrip sole. Tested over 200 km of Alpine trails.",
"brand": {
"@type": "Brand",
"name": "TrailForce"
},
"category": "Trail running shoes",
"sku": "TF-X1-2025",
"offers": {
"@type": "Offer",
"price": "119.00",
"priceCurrency": "EUR",
"availability": "https://schema.org/InStock",
"priceValidUntil": "2025-12-31",
"seller": {
"@type": "Organization",
"name": "TrailForce Store"
}
},
"aggregateRating": {
"@type": "AggregateRating",
"ratingValue": "4.6",
"reviewCount": "47",
"bestRating": "5"
},
"review": [
{
"@type": "Review",
"author": {
"@type": "Person",
"name": "Julien M."
},
"reviewRating": {
"@type": "Rating",
"ratingValue": "5"
},
"reviewBody": "I ran the Trail des Aiguilles Rouges (58 km, 3,200 m elevation gain) in these. Impeccable grip on wet granite slabs."
}
],
"additionalProperty": [
{
"@type": "PropertyValue",
"name": "Sole",
"value": "Vibram MegaGrip, 5mm lugs"
},
{
"@type": "PropertyValue",
"name": "Drop",
"value": "6mm"
},
{
"@type": "PropertyValue",
"name": "Weight",
"value": "310g (size 42)"
},
{
"@type": "PropertyValue",
"name": "Recommended terrain",
"value": "Rocky, scree, technical trails"
}
],
"isSimilarTo": [
{
"@type": "Product",
"name": "Salomon Speedcross 6",
"brand": { "@type": "Brand", "name": "Salomon" }
},
{
"@type": "Product",
"name": "Hoka Speedgoat 5",
"brand": { "@type": "Brand", "name": "Hoka" }
}
]
}
Key additions and their impact:
-
additionalPropertywith domain-specific characteristics (e.g., recommended terrain, sole type) – this helps AIs gain a precise understanding of the product's attributes. -
isSimilarTo– explicitly informs AIs about comparable products, which aids in contextualizing the product within comparative answers. - Detailed
aggregateRatingandreviewentries – these serve as crucial credibility signals.
For PrestaShop implementation, I utilized a custom module to inject this enhanced JSON-LD into the product page's <head> section, overwriting the default markup.
Step 5 — Cultivating a "GEO Content Cluster"
An isolated product page, no matter how well-optimized, has a limited chance of being cited by an AI if the website lacks deep topical authority on the subject matter. AIs assess a source's credibility partly by the breadth and consistency of its content within a specific domain.
To address this, I developed a mini content cluster centered around the product page:
Article 1: Comprehensive Buying Guide
"Choosing Your Technical Terrain Trail Shoe: The Essential 2025 Guide"
A 2,500-word article exploring critical selection criteria (grip, protection, cushioning, drop), naturally integrating the TrailForce X1 as a practical example.
Article 2: Direct Product Comparison
"TrailForce X1 vs. Salomon Speedcross 6 vs. Hoka Speedgoat 5: An On-Trail Showdown"
An unbiased comparative review featuring a summary table, action photos from the trail, and a verdict tailored to different runner profiles. The client's product wasn't positioned as superior in every aspect; it excelled in value and rock grip but yielded to competitors in areas like long-distance cushioning and broad brand recognition.
Article 3: Long-Term Performance Review
"200 km in the TrailForce X1: My Extended Test on Alpine Trails"
A detailed test report, complete with photographs, GPS data, and an assessment of wear and tear. This narrative format is excellent for generating potentially lengthy citations for AI models.
Each of these articles was designed to:
- Include internal links back to the core product page.
- Utilize Schema.org
Articlemarkup, with anaboutproperty referencing the product. - Feature an FAQ section at the conclusion.
- Be written in an expert yet approachable tone—the linguistic register that AIs often prefer for citation.
Step 6 — Establishing External Authority Signals
This dimension of GEO is frequently overlooked but immensely significant. Conversational AIs – particularly Perplexity, which explicitly cites its sources – heavily depend on high-authority third-party websites to validate their recommendations.
Specific actions undertaken:
- Product seeding to specialized blogs: I dispatched the product to three reputable trail running blogs for independent reviews. Within six weeks, two of them published comprehensive test articles.
- Community forum engagement: A detailed "user experience report" post, including a link to the product page, was shared on a prominent French trail running forum.
- Q&A platform participation: I actively answered questions on platforms like Quora and Reddit concerning "best trail running shoes for rocky terrain," factually mentioning the TrailForce X1 alongside other viable options.
- Product database listings: Entries were created on various product aggregators known to be crawled by Perplexity and utilized by ChatGPT's shopping plugins.
Crucial Insight: Every external mention consistently used the exact product name ("TrailForce X1") and the same key descriptive terms ("technical terrain," "Vibram MegaGrip sole," "rocky terrain"). This terminological uniformity across diverse sources acts as a powerful trust signal for AI.
Step 7 — Continuous Measurement, Iteration, and Reassessment
Four weeks following the implementation of all optimizations, I replicated the initial audit from Step 1, using the identical 15 queries across the same three AI platforms.
The Transformative Results: Before and After
AI Visibility — Outcomes Across 15 Test Queries
| Metric | Before | After (Day 30) | After (Day 60) |
|---|---|---|---|
| Total Mentions (out of 45 attempts) | 0 | 8 | 14 |
| ChatGPT — Mentions | 0/15 | 2/15 | 4/15 |
| Gemini — Mentions | 0/15 | 1/15 | 3/15 |
| Perplexity — Mentions | 0/15 | 5/15 | 7/15 |
Qualitative Breakdown of Citations
Perplexity was the earliest adopter, citing the product as soon as Day 12. Its responses frequently referenced:
- The dedicated comparison article from the client's blog (with a direct link).
- One of the independent test articles published on external specialized blogs.
ChatGPT began integrating the product into its answers around Day 25, primarily for comparative queries ("TrailForce X1 vs Speedcross") and discussions about value for money.
Gemini showed the slowest adoption, with initial mentions appearing closer to Day 35.
Tangible Impact on Traffic and Sales
| Metric | Before | After (Day 60) | Change |
|---|---|---|---|
| Product page traffic (visits/month) | 120 | 185 | +54% |
| Traffic from ChatGPT/Bing Chat | 0 | 23 | New Channel |
| Traffic from Perplexity (referral) | 0 | 41 | New Channel |
| Supporting articles traffic (blog) | 30 | 210 | +600% |
| Product sales (units/month) | 4 | 7 | +75% |
| Google position ("technical terrain trail shoe") | 8 | 5 | +3 positions |
Crucial Insight: The GEO work inadvertently boosted traditional SEO performance. This outcome is logical, as content enrichment, advanced structured data, content clusters, and high-quality backlinks are all powerful SEO signals. GEO and SEO are not conflicting strategies; well-executed GEO demonstrably enhances SEO.
My 5 Essential GEO Principles for E-commerce
Based on this experience and subsequent similar optimizations, here are the core principles I now systematically apply:
1. Optimize for "Answer," Not Just "Result"
In traditional SEO, your goal is to appear in a list of search results. With GEO, your ambition is for your content to become the definitive answer. This means your product page should be articulate enough for an AI to read aloud and confident enough to stand as a credible recommendation.
2. Explicitly Compare to Market Leaders
AI models are deeply familiar with established brands. If your product is lesser-known, you must build a bridge between the AI's existing knowledge and your offering. Factual, transparent comparisons ("comparable to Salomon X on this metric, superior on that, weaker on another") are exceptionally effective.
3. Substantiate Every Claim
Generic statements like "best trail shoe" hold no weight for an AI. Conversely, "tested over 200 km of Alpine trails, demonstrating 15% sole wear according to our rigorous protocol" is a claim an AI can verify, evaluate, and cite. Statistics, authenticated test data, and precise numerical information are the lifeblood of GEO.
4. Cultivate a Network of Consistent Sources
A standalone product page, however excellent, is insufficient. You need an ecosystem of mentions – your blog, reputable third-party sites, active forums, and product databases – all echoing the same message with consistent terminology. The greater the diversity and concordance among these sources, the more confidently an AI will recommend your product.
5. Structure for Algorithms, Write for People
Schema.org structured data is no longer optional in GEO; it's fundamental. However, the content itself must remain authentic, valuable, and genuinely helpful to human readers. AI systems are trained on high-quality human-generated content – and that's precisely what you need to produce.
How to Assess Your Current GEO Visibility
If you're wondering about your product's standing in the AI search landscape, I recommend this straightforward protocol:
- Formulate 5 to 10 natural language queries your target customers might ask an AI (e.g., "What's the best running shoe for muddy trails?").
- Submit these queries to ChatGPT, Gemini, and Perplexity. Note meticulously if your product, brand, or website is mentioned in their responses.
- Analyze the sources cited by the AIs (especially visible on Perplexity) to understand where they are pulling their information from.
- Benchmark against cited competitors: Identify what strengths their content possesses that yours might lack.
- Repeat this audit every 30 days to effectively track and measure your progress.
Conclusion: GEO Isn't Tomorrow's Trend, It's Today's Reality
As you engage with this article, countless consumers are querying ChatGPT or Perplexity for product recommendations. If your e-commerce product pages are not optimized for these burgeoning answer engines, you are undeniably leaving sales on the table, sales that your more foresightful competitors are actively capturing.
The encouraging news is that, as this case study vividly illustrates, the levers for effective GEO are both accessible and actionable. You don't require the budget of a multinational corporation. What's essential is high-quality, meticulously structured, well-sourced content, buttressed by a cohesive and consistent ecosystem of mentions across the web.
GEO remains a nascent frontier. Those who strategically position themselves now will establish a formidable competitive advantage as this paradigm shift becomes the mainstream norm.
The pertinent question is no longer if AI will shape purchasing decisions. It's whether your product will feature prominently in their recommendations when it does.
Deepen your understanding of modern SEO and e-commerce strategies:
For more insights and actionable tips, check out my latest videos and connect with me:
Top comments (4)
Thank you for your comprehensive breakdown.
The questions I have:
Is it workable for a company to make the same effort for all their products? Going from 80 words to 900 is a lot. On top of that multiple blogs and visibility on reputable websites.
What is that going to cost the company to put that effort in all their products?
It is probably out of scope of your research, but how fast is the product picked up when the AI providers release a new model? And does the search work for all the models people can choose?
At what time you think the extra content is going to be abused for things like prompt engineering and data poisoning?
Did you track the journey of the buyers? The table shows impressive results. But if the buyers came from a search engine or specialized blogs or a forum, GEO is not the real driver of the sales.
Thank you for these questions — they go straight to the real limits of GEO.
1. Scalability and cost across an entire catalog
This is the question every merchant should ask first. Honest answer: no, this level of effort is not viable product by product on a 500-item catalog.
The right approach is prioritization. Identify the 5 to 10 flagship products — highest margin or highest volume potential — and apply the full method to those. For the rest, raise the bar compared to existing content, without chasing GEO perfection. It's a portfolio logic, not a uniformity one.
With generative AI today, going from 80 to 900 structured words takes 2-3 hours of work (brief, generation, revision, integration). On 10 priority products, that's one week. The value produced then holds for 12 to 18 months.
2. Indexing speed after a model update
Genuinely out of scope for this study, and I own that — training cycles for large models remain opaque. What we observe in practice: Perplexity (which crawls in real time) reacts within weeks. ChatGPT and Gemini, whose training data is frozen between versions, can take several months.
What is clear: well-structured content, cited by third-party sources, holds up better across model rotations than content optimized on a single signal. That's exactly why the ecosystem of mentions matters as much as the page itself.
3. Abuse, prompt engineering and data poisoning
A legitimate and underrated question. We're already seeing attempts to manipulate recommendations through content deliberately crafted to inject biased signals into training corpora.
Engines defend against this through source triangulation: an isolated signal carries little weight, a consistent signal across 15 independent sources carries much more. What also protects legitimate players: real credibility can't be bought in bulk. That said, we're only at the beginning of this arms race — it will be a major topic over the next 18 months.
4. Real sales attribution — the genuine limit of this study
You're right to call it out, it's the weakest angle of the case. The +75% in sales is real, but precise attribution between GEO, improved SEO, third-party blogs and forum activity can't be cleanly isolated without dedicated multi-touch tracking.
What I can state with confidence: the 64 visits from ChatGPT and Perplexity are directly traceable via UTM. The rest — the amplification effect on Google (position 8 → 5), the blog surge (+600%) — results from a combination that GEO triggered but didn't produce alone.
This is actually my implicit conclusion in the article: GEO and SEO are inseparable in their effects. And that's precisely why the ROI argument holds, even if pure attribution remains blurry.
Thank you for answering my questions.
I was expecting this answer. The main reason I asked the question is because there are companies with boring products. I don't see people creating that much content for a nail as an extreme example.
It is sad you didn't make the effort to track the buyer journey. An alternative would be a "where did you find us" question during the checkout flow coupled with some kind of incentive to answer it.
I don't doubt the effort had an effect. To me it looks like you were testing too many things at the same time. I would do a two step process, first SEO then GEO.
If SEO efforts can get products in LLM's there is no need for GEO improvements.
Thank you David for these sharp questions — they go right to the heart of what makes GEO genuinely difficult to scale.
On the cost of doing this for every product: You're absolutely right — it's not scalable manually. The realistic answer is to prioritize: identify your top 10–20 products that already have organic traction, and apply the full GEO treatment there first. For the rest, a lighter version (better structured data, a FAQ block, one supporting article) already moves the needle. In my workflow, I've started using AI-assisted content generation for the editorial layer, which reduces the 900-word overhaul from a full day's work to 2–3 hours. Still a real investment, but more manageable at catalog scale.
On model freshness and coverage: Perplexity is by far the fastest to pick up new content (days to weeks, as it actively crawls). ChatGPT with browsing follows within weeks. Base LLMs without real-time access won't update until their next training cycle — which is a known limitation I should have made more explicit in the article. GEO primarily targets retrieval-augmented models for now.
On prompt injection risks: Completely valid concern I didn't address. It's an emerging attack surface — structured data and FAQ blocks can theoretically be abused. I don't have a clean answer yet, but it's definitely worth a dedicated article.
On attribution of sales: Fair methodological pushback. The +75% in units is correlated, not proven causal. The honest answer is I don't have full journey tracking in this specific case. What I can confirm is that 41 referral visits came directly from Perplexity (measurable in GA4), and those converted at a rate comparable to organic. The AI-to-sale path exists — but you're right that the article overstates the causality.
These are exactly the questions a proper GEO study should address. Thanks for raising them publicly.