Glossary · by Source of You
The words for how AI reads you.
When people ask an assistant about a name, it answers from whatever it can find. These are the terms for making the truth easy to find, read, and cite: the plumbing engines read, the signals they weigh, and the pieces Source of You builds from one record.
01
The answer era
How search changed
Buyers and recruiters increasingly ask an assistant a question instead of scanning a list of links. These are the words for that shift.
- Answer engine
- A system that answers a question directly (ChatGPT, Claude, Perplexity, Google AI) instead of returning a page of links.
- AEO (Answer Engine Optimization)
- Structuring true information so AI assistants can read, trust, and cite it. The AI-era counterpart of SEO: legibility instead of tricks. It is the term we use.
- GEO (Generative Engine Optimization)
- Another name for the same shift, framed around the whole entity rather than a single page. Most people use GEO and AEO interchangeably; so do we.
- SEO
- Search engine optimization: earning a rank in a list of blue links. Good SEO foundations still help, but an answer engine names a few sources instead of ranking ten.
- Answer fuel
- Clear, self-contained passages an engine can lift into an answer without rewriting: a direct definition, a real Q&A, a stated fact. Vague or purely promotional copy gives an engine nothing to use.
02
What engines read
The plumbing of being legible
Before an engine can use a page, it has to reach it, parse it, and know who it is about. These are the machine-readable pieces that make that possible.
- Canonical URL
- The one address that counts as the official copy of a page. Everything else should point at it, so an engine cites one place instead of guessing among duplicates.
- JSON-LD
- Machine-readable facts embedded in a page: name, role, links. The format engines parse for identity. It is the difference between being parsed and being skimmed.
- Person / ProfilePage schema
- The specific JSON-LD types that tell an engine a page is about a person and states who, instead of leaving it to infer identity from prose.
- FAQPage schema
- Structured question-and-answer markup an engine can quote verbatim. The machine-readable form of answer fuel.
- sameAs
- A list of other profiles and pages that refer to the same person, so an engine can connect a scattered identity into one recognized entity.
- robots.txt
- A file at the root of a site that tells crawlers which paths they may read. It is a URL anyone can fetch, so what it permits is a checkable fact.
- llms.txt
- A plain-text index that tells AI systems what a site contains and which pages speak for whom. A handful of tools and agents read it.
- AI crawler
- The named agents answer engines send to read the web, e.g. GPTBot, ClaudeBot, PerplexityBot, Google-Extended. A site can allow or block each one in robots.txt.
03
What engines weigh
The four signals we score
When Source of You reads how answer-engine-ready your pages are, it scores four crawlable signals, each 0 to 25, into a single 0-100 number. These are the axes, and what each one asks.
- Findable
- Whether a clear, current page you own names you and states who you are, so your name resolves to something you control instead of a stale profile or someone else entirely.
- Structured
- Whether Person and ProfilePage structured data plus a canonical link are present, so engines read your identity instead of guessing it from prose.
- Corroborated
- Whether more than one page names you and a sameAs graph links to profiles and proof you do not fully control. Self-assertion reads as marketing; corroboration reads as fact.
- AI-legible
- Whether the pages that name you are actually crawlable by the AI engines, with FAQ answer-fuel they can quote.
- E-E-A-T
- Experience, Expertise, Authoritativeness, Trust: the general idea that a page proves who stands behind it (real people, credentials, contact, provenance). Recommending anyone is a trust decision.
04
Source of You
The product vocabulary
The terms specific to how Source of You works, defined the way the product actually uses them.
- Professional Identity Record (PIR)
- The single, self-owned, structured record that is your source of truth. Every surface is generated from it.
- One record, every output
- The thesis. You build one record; the product renders every surface from it (source page, resume, LinkedIn summary, bio, audit). Change the record once and every render updates.
- Source page
- Your flagship public surface: one crawlable, canonical, structured page that states who you are, at your handle. Publishing it is a Pro feature.
- Record vs output
- The record is the source of truth. An output is a lossy render of it for a particular surface. The record is what you edit; outputs are what you share.
- Grounding
- The deterministic gate that checks generated text against your record. A figure or claim it cannot find there is dropped rather than published. Honesty is enforced here, not just promised.
- Provenance
- Every surfaced claim traces to something you entered or clearly-badged record-derived text. Provenance over guesswork, structure over scraping.
- Visibility audit
- A read of how discoverable and legible you are to AI: a deterministic 0-100 signal score you can run free, and a deeper, sourced review of how the open web describes you (a Pro output).
- Source agent
- The ask loop on a published page. It answers only from your record, declines what it cannot ground, and passes those unanswered questions back to you. It never fabricates.
- YouSourced
- The jobs layer: a free public feed straight from company boards, and (for Pro) apply-packs generated from your record.
See where you stand
Test how findable, structured, corroborated, and AI-legible your identity is today, with the four signals above. Free, no account. Then build the one record every surface is generated from.