AI Foundations

How AI systems find, trust, and recommend you.

AI visibility isn’t one problem. It’s a stack. Machines have to identify you before they can trust you, and trust you before they’ll recommend you.

These are the foundations, in the order they build on each other.

Start the Sequence See the Whole Map

By Sorilbran StoneVisibility Engineer · Founder, Five-Talent Strategy House

The three visibility conditions

01LegibilityCan the machine tell who you are?

02EligibilityDoes it trust you enough to include you?

03ExtractabilityCan it safely pull and cite what you’ve published?

Miss one, and the other two can’t carry you.

4 articlesRead in sequence

About an hourStart to finish

For practitionersFounders, marketers, and operators

Still growingNew pieces get added here

Who this is for

You’re doing the work. The results aren’t consistent.

This hub is for founders, marketers, and practitioners whose AI visibility efforts aren’t producing consistent results. Most of the time, it’s not effort. It’s which layer you’re working on.

Start with the map

LLM Behavior Optimization: A Mental Model

The whole system in one read. How LLMs actually choose, why SEO playbooks miss, and the diagnostic for finding which layer is breaking. Every article below zooms in on one part of this map.

Read the Mental Model →

01The three LLM behaviors
02The three retrieval failures
03The three visibility conditions
04How the behaviors and conditions connect
05The diagnostic

Read in sequence

Four articles. Each one builds on the last.

Identity first. Then what goes wrong when machines go looking. Then how they actually behave. Then where you’re trying to land.

Key terms

The language of this series.

Quick definitions for the terms you’ll run into across these articles. Each one is unpacked in full in the article it comes from.

Canonical bio01: One authoritative, current statement of who you are, on a page you own, that every other profile can agree with.
Identity regression01: When AI systems fall back to an older, outdated version of you after a key source disappears.
Mirroring02: When a machine can’t reach your page and answers from other sites that covered you instead.
Recall03: What the model already knows about you from its training data.
Retrieval03: What the model looks up live from the web or connected sources.
Inference03: What the model assumes it already knows. It answers without looking, often confidently, with no warning.
Cadence Gap03: The lag between training cycles, when new signals about you exist but haven’t been learned yet.
Legibility04: How reliably a machine can recognize you as one coherent entity and describe you accurately.
Eligibility04: Whether a machine judges you credible enough on a topic to include you in an answer.
Extractability04: Whether a machine can safely pull usable, citable content from what you’ve published.
Minimum Viable Knowledge Graph: Six nodes machines need to describe, match, and recommend you: Identity, Specialization, Audience + Context, Proof, Expertise, Connection.
The full glossary: See the AI Visibility Glossary →

Go deeper

When you’re ready to build.

The foundations explain how machines behave. These turn that into structure you can actually build.

The blueprint · 8 min The Minimum Viable Knowledge Graph

The six nodes machines need, and how to seed the connections between them.

Read → Framework library The Frameworks

Entity Archaeology, Hedge Signals, Compensating Signals, Blue Puddles, Load-Bearing Nodes, and more. Built in practice, not theory.

See the library → Book · $14.99 UnInvisible

An AI Visibility Playbook. The MVKG framework and the fundamentals, in one read.

Get the book → Notion tool · $79 Signal Builder

Score your signals, work through the build actions, and track your progress.

Get Signal Builder →

This section is still growing. New thinking shows up first on LinkedIn, then lands here.

Follow on LinkedIn ↗

Curious where you stand?

Which layer is breaking for your business?

Visibility Intelligence is how I look at a real business through all three conditions: how search and AI systems actually see it, where the trail goes cold, and where it has room to win.

Take a look insideWhat it examines, how it works, and what real cases turned up.

See How It Works →

Or keep reading. Start with the canonical bio.