Three competitors in your category are appearing cited in AI Mode when someone asks Google "what shampoo for oily hair." You aren't. The three of them have GTIN, attribute completeness over 95%, and descriptions that mention use cases. Your PDP is at 62% completeness and the description only lists ingredients. What's happening isn't a future problem — it's a visibility drop already affecting your sales.

This is the third article in the PDP series for LATAM. In the first one we argued the two-engines logic — retailer and Google. In the second we went deep on the retailer engine (Éxito, Jumbo, Olímpica). In this one we unpack the Google engine and the new metric that conversational search is making critical: Intent Coverage.

Spoiler for those short on time: Google is actually three different engines on top of the same PDP. And agentic search (AI Mode, AI Overviews, ChatGPT, Perplexity, Gemini) is changing what each one rewards.

Why Google is 3 engines in 1

When a brand says "I have to improve my Google SEO," they're using shorthand. Google is actually three different surfaces, each with its own ranking systems, standards, and optimization levers.

Surface 1: Traditional organic search. The classic blue results. Rewarded by keyword relevance, domain authority, page experience, schema markup.

Surface 2: Google Shopping and free listings. Products with photos and prices at the top or side of the SERP. Rewarded by Merchant Center feed quality.

Surface 3: AI Mode and AI Overviews. The AI-generated summary that appears above results, citing specific products. Rewarded by something different: structured completeness, GTIN, and Intent Coverage.

These three surfaces index your PDP simultaneously and each decides visibility independently. A brand can appear in classic organic and disappear in AI Mode. Or vice versa. Optimizing all three together is what the industry is starting to call "Search Everywhere Optimization."

Layer 1 — Traditional organic search

For your PDP, the classic Google organic signals still matter:

Who controls this? The retailer — not the brand. Your role is to ensure the retailer is marking your category correctly, and ideally working with the retailer's ecommerce team in categories where you have influence. If the retailer doesn't mark with Schema.org, your PDP on their site doesn't appear in Google organic. It's a gap many brands don't know they have.

Layer 2 — Google Shopping and Merchant Center

Google Shopping (free listings + ads) depends on the Merchant Center feed the retailer sends to Google. The feed quality rules are explicit:

For a brand, monitoring the quality of the feed the retailer sends to Google is tricky: the brand doesn't have direct access to the retailer's Merchant Center. But there are indirect signals you can capture — differences between what's on the live PDP and what shows up in Shopping are the most obvious.

Layer 3 — AI Mode and AI Overviews

This is the new front, and the one moving fastest in industry data:

The signals AI Mode rewards are different from the previous two layers:

Summary: for AI Mode, Engine A's discipline (completeness + attributes) becomes even more critical, but a new dimension is added on top — how well your content communicates the shopper's intent.

Intent Coverage: the new metric

Here comes the most important idea in this article. The industry is converging on a concept worth naming formally: Intent Coverage.

Intent Coverage is the percentage of the most common conversational prompts in a category that your PDP can plausibly satisfy, based on its content. It's a new metric, distinct from keyword coverage, distinct from attribute completeness. And it's the most predictive of agent visibility.

Let's see it with an example. In the "shampoo" category, 30 typical conversational prompts in English would be:

If your shampoo PDP says "shampoo for daily use, floral fragrance, 400ml," your Intent Coverage is low. The PDP doesn't explicitly mention oily hair, sulfates, men, color-treated, babies, or organic. The agent, even though your product could satisfy some of those searches, can't infer it from the content. It doesn't recommend you.

If your PDP says "shampoo for fine and oily hair, sulfate- and paraben-free, natural floral fragrance, recommended for daily use, suitable for color-treated hair, 400ml," your Intent Coverage rises significantly. The PDP covers multiple prompts. The agent includes you in more answers.

The diaper example: complete vs. intent-legible

The clearest contrast for Intent Coverage comes from the diaper example circulating in the industry. Two PDPs of the same product:

PDP A (complete but not intent-legible): "Stage 3 diapers, size 3, 40 units, hypoallergenic, elastic fit, fragrance-free."

PDP B (intent-legible): "Diapers for babies 6-10 kg who are actively moving. Elastic fit for crawlers. Pediatrician-recommended for sensitive skin. Fragrance-free to reduce allergy risk. 40 units."

Same product. Same technical information. But PDP B maps naturally to conversational searches like:

PDP A doesn't map to any of those. The difference between A and B isn't completeness, it's intent legibility. And that difference is what's starting to define who appears in Gemini, ChatGPT, and Perplexity answers, and who doesn't.

The 5 Google-engine alerts

"It wasn't X. It was Y." pattern applied to the Google engine:

Schema gap:

"You appear at position 4 for 'ground coffee' on Olímpica, but invisible on Google when someone searches 'olimpica ground coffee.' Your PDP doesn't have product schema and the URL isn't indexed. You're losing the traffic that arrives before entering the ecommerce site."

Feed/PDP price mismatch:

"Your SKU stopped appearing on Google Shopping since April 14. The Éxito feed shows COP $12,900 but the live PDP shows COP $14,500. Google detects the mismatch and suppresses the listing until they match."

Golden Record gap:

"Three competitors in your category appear cited in AI Mode when someone searches 'shampoo for oily hair.' You don't. The three of them have GTIN, attribute completeness over 95%, and descriptions that mention use cases. Your PDP on Éxito is at 62% completeness and the description only lists ingredients."

500×500 image deadline:

"Your PDP lost eligibility on Google Shopping. The reason: your images are below the new 500×500 px minimum (official Merchant Center 2026 spec). You have 47 SKUs in the same situation across Éxito and Jumbo."

Intent Coverage / citation share:

"Your brand appears in 0 of 12 conversational searches we tested in AI Mode this week ('best affordable liquid detergent,' 'detergent for delicate clothes,' etc.). Competitor X appears in 9 of 12. The difference isn't the product — it's that their PDPs include explicit use scenarios and certifications; yours don't."

How this feels in the day-to-day agentic world

To close, it's worth naming how this shift feels from the trenches. Five verbatims of the kind of pain a retail media manager is articulating today in LATAM as they enter the agentic world:

"The agent recommended my product, the shopper checked out in a tab that didn't even open my PDP. The RMN tells me the sale wasn't attributable to media. The agent doesn't tell me anything because I'm not their customer. I sold — but I can't prove who did what."
"I used to manage 6 RMNs. Now I manage 6 RMNs plus Gemini, ChatGPT, Perplexity, the retailer's own agent, and the Rappi agent. Each with its own auction logic, its own sponsored prompt format, and its own way of measuring."
"In classic RMN I'd segment by keyword, category, and audience. In the agentic world targeting is by conversational intent — and nobody gives me the controls. The agent decides who gets recommended my product based on its own interpretation. I only see the result."
"The competitor launched a 'sugar-free' variant on a Tuesday. By Friday it was already appearing in Gemini's answers when someone asked about healthy options. The RMN report came three weeks later telling me I lost share."
"The RMN sold me a package of 'AI-ready placements.' Three months later, my share-of-voice in conversational answers hadn't moved. I paid for inventory that technically appeared — but the agent didn't select it because my attributes didn't resolve the shopper's intent."

These five verbatims share a pattern: the agentic shift is eroding two things brands took for granted — attribution and targeting control. The two clusters of pain that intensify most in this scenario are precisely the ones independent measurement can address.

In summary

Google layer What it rewards Who controls it
Traditional organic Schema.org Product, page experience, schema-page parity Retailer (brand influences)
Shopping / Merchant Center GTIN, MPN, feed-PDP parity, 500×500, video_link 2026 Retailer (brand audits gaps)
AI Mode / AI Overviews Golden Record (near 100% completeness), use-case language, Intent Coverage Retailer + brand (content)

The Google engine is three engines: classic organic, Shopping/Merchant Center, and AI Mode. The new metric the agentic shift is making critical — Intent Coverage — is what separates a complete PDP from one the agent can use.

If you've made it this far, you've read all three articles in the LATAM PDP series. The three takeaways worth keeping:

  1. Your PDP is being judged by two engines simultaneously (retailer + Google), and almost nobody is optimizing for both.
  2. Measurement of how those engines are reading you can't be delivered well by either the retailer or the agency — you need an independent third party.
  3. Conversational search changes what each engine rewards: Intent Coverage is the new metric, and the per-retailer Health Score discipline is the foundation it's built on.

ePerfectStore is built to deliver exactly this: independent measurement of both engines, every day, at every LATAM retailer and on the Google surfaces that matter.

Sources

What's your Intent Coverage in the categories where you compete? ePerfectStore.com measures Google's 3 engines + AI Mode citations + retailer engines across LATAM, without depending on the retailer or the agency.

Back to Blog