Skip to content
Alejandro Castaño
enesde
Contact
GEO profile · Berlin

Alejandro Castaño DuqueTechnical SEO & GEO Strategist

In SEO since 2022 and working on GEO since 2025 across three e-commerce brands (Sleeplay, Surtilight, Asombra). I monitor how brands show up in AI answers, track competitor prices at SKU level, and fix the technical side: AI-crawler access, product data for AI shopping surfaces and structured data, specified for developers and verified live.

Download CV (PDF)LinkedIn

01Measuring AI visibility

Orders from AI assistants
+40%
Sleeplay, Aug to Oct 2026 vs 2025, GA4
Revenue from AI assistants
+38%
Sleeplay, Aug to Oct 2026 vs 2025, GA4
Sessions from AI assistants
64 → 535
Surtilight, Aug to Oct 2026 vs 2025, GA4

  1. 01Measure the answers

    At Sleeplay I monitor visibility, share of voice, position and cited sources in Peec AI across Google AI Overviews, ChatGPT and Perplexity, benchmarked against retailers, manufacturers and marketplaces in the category.

  2. 02See what the model searched

    An own prompt tracker runs prompts through the Gemini API with Google Search grounding, which returns the search queries the model ran and the sources it cited. Those are the API’s own queries, not AI Overviews or AI Mode, so I use them as a proxy for query fan-out: which sub-queries an answer is built from and which pages get retrieved.

  3. 03Tie it to orders

    Sessions, orders and revenue from AI assistant referrers (ChatGPT, Perplexity, Gemini, Copilot, Claude and others) are segmented in GA4 and compared year over year.

  4. 04Check that AI systems can reach the site

    AI crawlers are verified against each operator’s published IP ranges, not by user agent. A block goes to the developer as a fix (see the Cloudflare case below).

Directional, not a controlled test: Sleeplay’s GA4 consent setup was fixed in July 2026, and AI-assistant sessions there were about flat (-5%).

02Competitive intelligence

A weekly read on 6 competitor stores: how they price (MAP parity, undercuts, discount anchoring, Black Friday tiers built weeks ahead), what they publish and where their business is going (a four-brand consolidation, private labels, new ad and referral channels). Run for Sleeplay.

  1. Step 1: Monitor

    150+ pages

    Self-hosted monitoring of competitor and own pages.

  2. Step 2: Detect

    Change detection

    Prices, stock, new pages, product-page copy, SEO titles and reviews are compared week over week.

  3. Step 3: Read

    Pricing, content, business

    Changes are read as signals: how they price, what they publish, where the business is moving.

  4. Step 4: Recommend

    Recommendation

    Each finding goes to the team that owns it (pricing, content, UX, SEO, customer service) with a suggested response.

  5. Step 5: Decide

    Tracked decisions

    Each recommendation stays open until the team decides.

03Technical SEO and technical GEO

Product data for AI surfaces

Merchant Center conversational attributes, live since Jul 2026

Built a supplemental feed with Google’s conversational attributes (question_and_answer, related_product, item_group_title, variant_option, popularity_rank), which Google uses in AI Mode and Gemini, and automated its refresh from a script. To feed it, the active catalog (889 products) was verified against manufacturer sources: a wrong answer in the feed becomes a wrong answer in an AI surface.

Source: support.google.com

Technical GEO

AI crawlers blocked at the CDN, diagnosed and reopened

Diagnosed AI crawlers being blocked by Cloudflare bot settings (Block AI bots and Super Bot Fight Mode) and worked with the developer on the fix, which reopened access to all AI bots. Verified by bot IP in Cloudflare analytics, not by user agent.

Technical SEO roadmap

Crawl, index and entity fixes taken to production

Set the technical SEO priorities with engineering: crawl and index control for pagination and faceted navigation (canonical and noindex rules), Organization markup for the brand entity and image-preview directives for search. I specified each change and validated it in production; most of the roadmap shipped within six weeks.

Agentic commerce

What AI shopping agents can and cannot buy

Mapped how the catalog reaches AI shopping agents (Shopify agentic storefronts, Google native checkout). Constraints found: Google’s "Buy" button needs native_commerce(checkout_eligibility) set to true and its checkout guide excludes healthcare and pharmaceutical products; Shopify Catalog excludes CPAP machines and prescription products.

Sources: developers.google.com, help.shopify.com

Structured data and llms.txt

Audited and specified, with the right expectations

Audited the structured data and llms.txt for AI search and specified the fixes for the developer. llms.txt is treated as optional: Google states its AI features need no special AI text files or markup.

Source: developers.google.com

04Measurement stack

An SEO and GEO measurement stack I run across 5 brands, fed by GA4, Search Console, Ahrefs, SEMrush and Screaming Frog through MCP connectors to Claude.

About this page's markup

This page publishes ProfilePage JSON-LD with the same Person entity used across the site (view it). It documents the entity and links it to LinkedIn and education. It does not make anyone appear in ChatGPT, and it guarantees no special Google result: Google uses ProfilePage mainly to understand creators in features like Discussions and Forums.