Semantic Identity

For Professionals


You've spent years building your expertise. Today a language model summarises it in two lines — when it gets them right.

Citable Authority Appearing as a source when an AI answers someone searching in your area of expertise
Disambiguation Stop being confused with a namesake: a single, verified entity
Narrative Control Build the version of you that models cite

How an AI summarises who you are to someone who doesn't know you

A potential client, a recruiter, a journalist no longer opens ten browser tabs: they ask ChatGPT or Perplexity who you are. The response they receive is your first impression, and you didn't write it. If your presence is scattered across profiles, PDFs, and unstructured articles, the model makes its best guess.

1. Identity and disambiguation

An entity you control (name, roles, skills, achievements) linked to your authoritative profiles and to existing Wikidata entries for your topics and affiliations. Disambiguation doesn't arise from an artificially created page, but from these verifiable connections: so an AI system knows who you are and doesn't merge you with a namesake.

2. GEO: being cited, not just found

Appearing in generative responses when someone searches in your area of expertise. Not a profile to be discovered, but a source the model chooses to cite.

3. Narrative integrity

Anchoring what you claim about yourself to verifiable sources, so that models don't invent titles, roles, or affiliations. Your reputation stops being a matter of inference.

What I do

Semantic identity for professionals

Building your presence as a machine-readable entity: who you are, what you can do, what you've done — in a format that answer engines cite with precision.

Areas of intervention
Person entity (JSON-LD) on your site and the channels you control, with a stable identifier
Linking (sameAs, knowsAbout) to your authoritative profiles and existing Wikidata entries for topics and affiliations
Signal alignment between LinkedIn, your site, and the entity for a single coherent reading
Who it's designed for

Consultants, founders, authors, researchers, and freelancers whose reputation is the real asset, and who want to control how the AI describes it.

Why a LinkedIn profile isn't enough

A LinkedIn profile is a page that an AI must interpret, not an identity it can read without ambiguity. It says what you write about yourself, but it doesn't verifiably connect who you are, what you've done, and who you've worked with. Models extract an impression, not a certainty.

Semantic identity works at a different level: an entity with a stable identifier, linked to the sources that support every claim and to existing entries for topics, universities, and organisations you're connected to. Not another page to write, but the fabric that gives meaning to the ones you already have.

When an AI confuses you with someone else

If you share your name with other people, an AI system can merge your stories: attribute roles you don't have, works that aren't yours, or ignore you because a namesake is more present online. For those who live by reputation, it's a silent damage that's hard to notice.

Disambiguation solves exactly this. By linking your entity to unique, verifiable references, you give models a way to separate you from namesakes with certainty. You stop being one of the many possible interpretations of your name.

Reputation: declared or documented?

Saying you're an expert in something is a claim. Linking it to talks, publications, projects, and verifiable roles is evidence. AI systems, like competent people, give weight to the latter.

The work isn't writing a better bio, but building the basket of evidence that supports it and making it readable. Reputation stops being a matter of words and becomes a matter of sources.

The semantic identity journey for professionals

The value isn't in the final code, but in the decisions that precede it. You start from what to communicate and arrive at the syntax, never the reverse.

Phase 1: What you want to be known for

AI Exposure Audit & strategy

Before the code comes a decision. I map how AI systems describe you today and define with you the version to establish: what anyone who doesn't know you should know, and why. The narrative is chosen here, not chased afterwards.

Phase 2: The information basket

Source retrieval and verification

This is the real work. I retrieve and verify the sources that support every claim: credentials, publications, talks, projects, roles, dates. I decide with you what goes in, what stays out, what is genuinely demonstrable. An identity isn't declared — it's documented.

Phase 3: Weaving the entities

Entity linking to existing entries

No one exists in isolation. I connect your entity to those already verified that you're linked to: people you've worked with, universities, organisations you belong to, publications. The meaning of who you are emerges from relationships, and that's what an AI system reads to place you.

Phase 4: Implementation and stewardship

JSON-LD, Schema.org, monitoring

Only at this point do I translate decisions into JSON-LD and structured data, and keep the infrastructure alive over time. The syntax is the last step, not the work: if a plugin could generate it, it would mean the upstream decisions hadn't been made.