AI Search Audit: diveshop.pt — 19/100 & 18 Critical Blockers

I picked diveshop.pt for this audit because it is a local business in my market — Portugal. I live here, I know the diving scene, and I wanted to see what my scanner would find on a real e-commerce site selling high-end gear like rebreathers and regulators.

The result: 19 out of 100. Zero citations across 30 Perplexity queries. Eighteen specific issues flagged. Here is what my tool found, what it means, and — honestly — what fixing these things would and would not achieve.

What the Scanner Reported

These numbers are from an actual scan I ran through getaisearchscore.com. The crawler hit their sitemap, pulled product pages, checked for structured data, and then ran 30 relevant queries through Perplexity to see if diveshop.pt got cited anywhere. It did not.

The Three Biggest Problems

My tool flagged 18 issues across four dimensions. But three stood out as the most severe technical gaps.

1. No Products in the Sitemap

The sitemap index had categories, brands, and tags — but zero product URLs. If an AI crawler is looking for a Mares regulator or an Aqualung mask, it will not find them in the sitemap. This is a WooCommerce configuration issue, and it is surprisingly common.

Fixing this would remove a discovery barrier. AI crawlers could actually find the product pages. But discovery alone does not guarantee citation — the content on those pages still needs to be relevant to the question a user is asking.

2. Zero Structured Data

None of the scanned product pages had Schema.org markup. No Product schema, no Offer with price and availability, no LocalBusiness. To any machine reading this site, the pages look like unstructured text blocks rather than commerce data with prices, stock status, and product identifiers.

Adding JSON-LD would make the data machine-parseable. That is a real technical improvement. Whether it leads to more citations is a separate question — my own research on 441 domains found essentially zero correlation (r=0.009) between structural readiness scores and actual citation rates.

3. Mixed Language Signals

This one was interesting. The site UI — navigation, About page, Contact — is in Portuguese. But product descriptions are in English, straight from the manufacturer. The site sends contradictory language signals: is this a Portuguese resource or an English one?

For a Portuguese dive shop, the content language matters a lot. When someone asks Perplexity a question in Portuguese about dive gear in Portugal, the AI needs to find Portuguese-language content that directly addresses that question. English manufacturer copy does not do that.

Who Gets Cited Instead

I tested queries like "fato semi-seco vale a pena Portugal" (Is a semi-dry suit worth it in Portugal?). Perplexity cited Decathlon and JustDive. Not diveshop.pt.

Looking at what those competitors have, the pattern is clear:

• Answer-ready content: FAQ blocks and comparison tables that directly answer the question.
• Local context: They mention specific Portuguese dive sites and Atlantic water temperatures.
• Review signals: AggregateRating schema showing real customer feedback.

This is the part that matters most. My later research across 441 domains showed that content relevance is roughly 62x more important than structural readiness for getting cited. A page that directly answers the user's question in the right language, with local context, will beat a perfectly structured page that talks about something else.

What Fixing These Issues Would Actually Do

I want to be honest about what these fixes achieve. My tool identifies real technical problems. Fixing them removes barriers — your products become discoverable, your data becomes parseable, your language signals become consistent.

But removing barriers is not the same as guaranteeing citations. Think of it like opening a shop: fixing the sitemap is like putting up a sign. Adding schema is like putting price tags on products. Translating descriptions is like speaking the customer's language. All necessary. None sufficient on their own.

The thing that actually drives citations is content relevance — having pages that directly answer the questions people are asking AI. For diveshop.pt, that means Portuguese-language guides about diving in Portugal, comparison content, local buying advice. The structural fixes just make sure that content can be found and parsed when it exists.

The Fix List

Here is what I would prioritize, ordered by effort-to-impact:

Score Breakdown: Before vs. Projected After

Here is how the score would change across the four dimensions if the P0 and P1 fixes were implemented. These are projections from my scoring model, not measured results.

What I Learned From This Audit

Running my own tool on a site I know well was useful. The scanner correctly identified real problems — the missing sitemap entries, the absent schema, the language mismatch. These are genuine technical issues that any web developer would want to fix.

But this audit also reinforced what my later research confirmed: technical readiness is table stakes, not a competitive advantage. The sites that Perplexity actually cites for Portuguese diving queries have relevant, localized, answer-ready content. That is the hard part, and no audit tool can do it for you.

If you run a similar e-commerce site, fix the technical issues first — they are low-effort and remove real barriers. Then invest in content that directly answers the questions your customers are asking AI search engines. That is where the citations come from.

Related Resources

For the methodology behind the score, see the guide to scoring dimensions or the e-commerce checklist. For a comparison of available tools, see the 2026 tool comparison.