# Daniel Martínez > Daniel Martínez builds AI agents and websites. Born in Quito, he built his career in Medellín and now lives in Miami. He automated a fashion company with stores in Colombia and Ecuador with AI, from accounting to the warehouse, and documents in public what he learns about SEO for AI. https://danielmartinez.tech/en/ · hola@danielmartinez.tech · 2026-11-01 ## AI agents · Websites · SEO for AI I'm Daniel Martínez. I was born in Quito, built my career in Medellín and now live in Miami. I automated a fashion company with stores in Colombia and Ecuador with AI, from accounting to the warehouse. Here I share how I did it and document what I'm learning about SEO for AI. - 41 stores invoicing with the system I built. - 2 countries in operation: Colombia and Ecuador. - 6 years running a real company, area by area. - 1 in-house platform, from product to checkout. ## What I build ### AI agents AI that does the work, with real data and the company's own rules. - **Agents that take action.** A chat that doesn't just answer: it creates tasks, queries data, moves approvals and raises alerts, with each person's permissions. - **AI that reads documents and photos.** Invoices, cash closings, receipts, evidence. The AI reads, compares and flags only what doesn't add up. - **Voice and video with AI.** Assistants that speak with a natural voice, plus tutorials and videos generated from real material. - **Data and forecasting.** Python, statistics and machine learning to anticipate demand and decide ahead of time. ### Websites Fast, clear websites built to be found by Google and by AI. - **Pages that convert.** Clean, Apple-style design: one idea per screen, one clear button and real numbers that build trust. - **SEO for Google and for AI.** Fast sites, structured data and files that help ChatGPT, Claude, Gemini and Perplexity understand who you are. - **Online stores and Shopify.** Catalog, synced inventory and payments, so what sells online actually exists in the warehouse. - **Custom web apps.** Bookings, checkout, warehouse, internal portals. They run on a phone and don't need expensive servers. ## Projects ### An in-house platform for a fashion retail chain Colombia and Ecuador. I replaced the old software with our own platform: product master (PLM), warehouse (WMS), point of sale (POS) and a team portal. Today it invoices in 41 stores in Ecuador through the tax authority (SRI), connects to SAP Business One and, in Colombia, invoices through Siigo with the DIAN. ### A copilot that acts, not just answers Internal portal. A chat inside the company portal that creates and closes tasks, queries the database and moves approvals with each person's permissions. The screen and the agent use the same path, so neither falls behind. ### AI that reviews cash closings Stores in Ecuador. Each store uploads a photo of its closing report. The AI reads it and compares what was counted with what was sold; the area manager only hears about what doesn't add up. ### A voice assistant for management Management, Colombia. You ask out loud how sales are going or which expenses are waiting for approval, and it answers with a natural voice using real data. The AI only reads: the numbers come from the database. ### A training school with AI-made videos Internal training. Vertical tutorials to watch on a phone: the script, the illustrations and the voice are generated by AI from each lesson's text. A new lesson becomes a new video in minutes. ### A wholesale website for a factory Textile factory, Ecuador. An Apple-style page with video of the plant animated by AI from real photos, and a live quote builder that assembles the order and sends it over WhatsApp. ### A product page that assembles and comes apart Eyewear brand. Sunglasses that rotate and split into pieces as you scroll. No photo studio: the product views and the spin were generated with AI. ### Bookings for appointment-based businesses Medellín. A scheduling platform for barbershops and service businesses: no double bookings, client records and commission payouts. Each new business is configuration, not development. ### An app for weddings and events Colombia. Guests RSVP, see their table and upload photos; organizers manage tables, menus, vendors and check-in with NFC wristbands. ## My rules - **Build for real.** Real things that solve real problems. - **Truth over views.** No made-up numbers, no hype. - **Respect.** The company's data, and its clients' and suppliers' data, stays private. - **Always learning.** Every project leaves something the next one uses. - **See it coming.** Data and AI to anticipate, not to explain what already happened. ## Journey - **Quito · Data and statistics.** I studied data analysis with Python and statistics. That's where I fell in love with machine learning for demand forecasting. - **Medellín · at 24 · A company from scratch.** I moved to Colombia to open a fashion company in the country. - **6 years · The whole operation.** Store openings, logistics, accounting, HR. I worked in every area to keep the operation moving forward. - **AI first · An automated company.** We built an AI-first e-commerce and automated the company with AI, from accounting to logistics, on our own platform in Colombia and Ecuador. - **Today · Miami · Learning in public.** Today I live in Miami. I document what I've built and what I'm learning about SEO for AI, so it can help more people. ## FAQ **What is an AI agent, and how is it different from a chatbot?** A chatbot answers questions. An AI agent does things: it queries data, creates tasks, reads documents and flags what doesn't add up, always with the permissions of the person using it. **What is SEO for AI?** It's getting AI assistants (ChatGPT, Claude, Gemini, Perplexity) to find, understand and cite your information. It's similar to Google SEO, but adds new pieces: clear structured data, an llms.txt file, content that answers specific questions and mentions on other sites. **How do I get ChatGPT to recommend me?** There's no button. What works is being found (Bing, Google and IndexNow), being understood (a clear page with structured data and llms.txt) and being trusted (other sites describing you in the same words). In the lab I document every step with data. **Where do you live?** I was born in Quito, built my career in Medellín and now live in Miami. **What tools do you work with?** Claude, Gemini on Vertex AI, ElevenLabs and Veo for AI; Python for data; Cloudflare and Supabase for infrastructure; Shopify, SAP Business One, Siigo and electronic invoicing with Ecuador's SRI and Colombia's DIAN for integrations. ## SEO for AI Lab ### Day 0: how I set up my site so AI can find me (2026-11-01) https://danielmartinez.tech/en/lab/day-0-seo-for-ai/ I bought danielmartinez.tech on October 28, 2026. Zero pages, zero mentions. My goal: when someone asks an AI about AI agents or SEO for AI, my name shows up. This is the lab's first entry: what I did on day 0, and why. #### The idea in one line An AI recommends you when it can **find you**, **understand you** and **trust you**. Everything below tackles one of those three. #### 1. Being found - **Sitemap and Google Search Console.** I verified the site with an HTML tag, submitted the sitemap and requested indexing for the main pages. - **IndexNow.** A direct ping to Bing, which powers Copilot and is one of the sources behind ChatGPT search. It replied 202: accepted. - **Open doors for AI bots.** robots.txt lets in GPTBot, ClaudeBot, PerplexityBot, Google-Extended and others. I also checked Cloudflare, which can block AI bots: in my case all three types were set to “Allow”. #### 2. Being understood - **A clear page, in Spanish and English.** Who I am, what I build and where, with real cases and numbers I can back up. - **Structured data (schema.org).** Person, website, FAQ and articles in JSON-LD. Zero errors in the validator. - **llms.txt.** A plain-text summary designed for language models, plus a full version with all the content. - **Speed.** PageSpeed on mobile: performance 93, accessibility 100, best practices 100 and SEO 100. #### 3. Being trusted This is the part that's still missing, and the hardest: other sites talking about me in the same words. My name is very common, so I need the same photo, the same one-line bio and links between all my profiles so AI doesn't confuse me with another Daniel Martínez. #### What I'll measure - Every month, the same questions to ChatGPT, Claude, Gemini and Perplexity: do they name me? do they cite me? in what words? - Impressions and clicks in Google Search Console. - Which pages AI bots read. I'll publish the day 0 measurement in the next entry. #### What I learned today - AI finds you three ways: what it learned in training, what it searches live and what it reads when someone pastes a link. - Your site alone isn't enough: trust comes from outside. - Only publish what you can back up. A made-up number is a risk, not an advantage.