Every era of branding had a defining deliverable. The print era's was the logo. The identity
era's was the brand book — colors, type, voice, a hundred pages nobody opened twice. The digital
era's was the website. And the era we just entered has one too, though almost nobody is building
it yet: the entity file. Call it
entity.md — a plain, machine-readable document that tells every AI system who you are,
what you do, what you've proven, and how to represent you.
I've made the argument before that your brand is now a file AI reads. This essay is about the file itself — because "branding is evolving" stays abstract until you've seen the deliverable, and at my company we've stopped saying we build brands at all. We build extensive entity files.
What goes in an entity file?
Think of it as the canonical record a machine would need to describe you accurately and recommend you confidently:
- Canonical facts. Legal name, brands, locations, services, service areas, hours, licensing, leadership — stated once, correctly, in a form machines can quote instead of guess.
- Offers and proof. What you actually sell, and the verifiable evidence behind every claim: named partnerships, recognitions, real results. Claims without proof get discounted by AI systems the same way they should be by humans.
- The soul layer. Voice, values, positioning, the founder's actual thinking — written down so every AI that describes you (or writes for you) learns your soul instead of substituting a generic one. This is the "systems and soul" balance made literal: the entity file is the soul, encoded into a system.
- A question-and-answer corpus. The questions your customers actually ask, answered in your words — the raw material answer engines lift from.
- Terms for machines. Licensing, usage rights, and what AI systems may and may not do with your content. The machine era needs a machine-readable contract.
This isn't theory — you can read mine right now
Open jenniferbagley.com/entity.md — yes, the literal file. That is my entity record, live in production: identity, canonical facts, offers, proof, defined terms, relationships, and licensing terms for AI systems — one authoritative document built to be read by machines first. It sits alongside llms.txt (the curated index) and the full corpus, and every page here carries the connected structured-data graph underneath. Under it sits the structured-data graph connecting the person, the book, the company, and the speaking practice into one entity. When an AI system answers a question about me, it doesn't have to reconstruct me from fragments. I handed it the file.
Why the file beats the brand book
- It compounds. A brand book decays in a drawer. An entity file gets read by every crawler, assistant, and agent that encounters you — thousands of times a day, forever.
- It disambiguates. Machines confuse similarly-named businesses, outdated facts, and third-party guesses constantly. The entity file is your side of the story, stated authoritatively.
- It converts, invisibly. When buying moves through AI intermediaries — and I have that on the record — the businesses with clean entities get recommended, and the rest get summarized badly or skipped.
The evolution, named
Logo era: be recognized. Identity era: be consistent. Website era: be found. Entity era: be understood — by machines, correctly, everywhere. The companies that treat the entity file as a first-class deliverable — as the thing the brand actually is now — will spend the next decade being quoted accurately and recommended confidently while their competitors argue with a chatbot's hallucination of them.
The full manifesto of how we rebuilt our company around deliverables like this one — revenue engines, entity files, component marketplaces — is here. And if you want to know what an entity file would contain for your business, that's a conversation my team has every day.