--- title: "From Master Hair Academy to ENTIA" canonical: https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia language: en author: Fernando Vilches (https://entia.systems/#founder) publisher: ENTIA Systems (https://entia.systems/#organization) date_modified: 2026-09-17 alternate_language: https://entia.systems/fernando-vilches/notes/from-master-hair-academy-to-entia json_ld: https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia.json html: https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia --- # The demand was there. *The trust was not.* How a hairdressing academy became the laboratory that gave rise to ENTIA: verifiable identity infrastructure for machines. ## The first experiment was a real business Before ENTIA I built [Master Hair Academy](https://masterhair.academy), a training academy for hairdressing professionals with courses in balayage, barbering and hairstyling. I came from digital marketing and wanted a scalable, self-service operation: video content, recognisable mentors, support and a complete purchase funnel. My hypothesis was that generative search opened a window different from traditional SEO. A new brand could gain presence on a surface where domain age, ad spend and years of accumulated authority did not yet weigh the same. I built landing pages, topic clusters and structured representations, and started observing what different models returned. ## Citation was the wrong layer to experiment on I soon saw that a specific citation was not reproducible. The model changed, the wording of the question changed, the context and other parameters changed. I could observe an appearance, but not treat it as a stable variable. So I started modifying JSON-LD, page structures and URLs, and repeating queries. The method was deliberately empirical: change one variable, measure, break, measure again. I did not care whether a single answer improved; I wanted to find which signals survived consistently across different systems. > If I could not stabilise the citation, *I had to work one layer earlier:* retrieval and indexing of the entity. ## Commercial failure revealed the missing layer Master Hair Academy generated discovery, traffic and purchase intent. The problem appeared at the end of the funnel: customers wanted to check that the professional they admired was really tied to the course. They looked for external corroboration on the mentors’ official channels and could not find it. The product was discoverable, but a layer of identity and trust connecting the person, the course and the organisation in a verifiable way was missing. It was my first practical proof that **visibility is not identity, and identity is not trust**. ## From inference to layers of truth After the failure I kept working on the technical problem. I studied what happened when models had to fill in information from incomplete signals, and formulated what I later called **Cognitive Resistance**: the more stable, verifiable and coherent a representation is, the less room it leaves for a system to infer who the entity is. I turned entia.systems into my own laboratory. ENTIA’s representation appeared and disappeared as I changed the JSON-LD and the relationships between fields. Adding data was not enough: structure, coherence and the authority of the sources mattered. ## The source of truth became architecture In the first landing pages I had already seen that linking a claim to recognisable sources helped contextualise it. In ENTIA I took that much further: every relevant data point had to rest on a deterministic, traceable source. BORME, BOE, VIES and other official registries stopped being decorative links and became part of the identity model. And not every field is worth the same. An address, a tax identifier, a registry filing or an external mention should not weigh equally. That weighting work became the Risk Score and the rules that decide what goes into an Entia Home, and with what strength. ## The URL stopped being a landing page and became an entity For close to eight months I iterated on structures that performed unevenly across models and search engines. I was looking for a canonical representation stable enough for different systems to arrive at the same point, regardless of country or consumption format. Once that structure stabilised, the URL became the address of an entity: an Entia Home. From there the same identity could be served as HTML, JSON, Markdown, API and, later, MCP. > The method has not changed since the first laboratory - MEASURE - BREAK - MEASURE - OBSERVE - KEEP ONLY WHAT REPEATS Enquiries still reach [masterhair.academy](https://masterhair.academy) today, without me investing in it. I was not wrong about the approach: the demand existed and still exists. What did not exist was a way for a machine, and then a person, to check who was behind it. That missing piece is what I build at ENTIA. First with Codex and Claude on a website; today on an entire platform designed by me and executed by coding agents. ## Notes in this series - [From Master Hair Academy to ENTIA](https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia) · [JSON-LD](https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia.json) · [Markdown](https://entia.systems/en/fernando-vilches/notes/from-master-hair-academy-to-entia.md) - [Before you can be visible, a machine has to know who you are](https://entia.systems/en/fernando-vilches/notes/identity-before-visibility) · [JSON-LD](https://entia.systems/en/fernando-vilches/notes/identity-before-visibility.json) · [Markdown](https://entia.systems/en/fernando-vilches/notes/identity-before-visibility.md) - [I did not block the bots: I took them off the origin](https://entia.systems/en/fernando-vilches/notes/bots-at-the-edge-not-the-origin) · [JSON-LD](https://entia.systems/en/fernando-vilches/notes/bots-at-the-edge-not-the-origin.json) · [Markdown](https://entia.systems/en/fernando-vilches/notes/bots-at-the-edge-not-the-origin.md) - [Profile: Fernando Vilches](https://entia.systems/en/fernando-vilches) · [JSON-LD](https://entia.systems/en/fernando-vilches.json) · [Markdown](https://entia.systems/en/fernando-vilches.md)