The Visibility Code™ is a publishing and research initiative focused on Knowledge Engineering for Answer Engines—the emerging discipline of making publisher knowledge explicit, resolvable, attributable, and measurable in machine-mediated information systems.
It begins with a simple observation: the machine’s job changed. Search engines traditionally helped people find documents. Answer engines increasingly identify entities, retrieve facts, follow relationships, compare alternatives, and synthesize answers before a person ever reaches the original source.
That changes the publisher’s job, too.
Why The Visibility Code Exists
For decades, web publishing optimized the document. Publishers learned how to make pages crawlable, indexable, understandable, and discoverable.
But publishers frequently know more than their pages explicitly preserve. Identity, provenance, relationships, applicability, definitions, derivations, temporal scope, and other semantic structure may exist in databases, source files, application logic, and editorial systems only to disappear at the final publication boundary.
Human readers can often reconstruct that missing context. Machines may have to infer it.
The Visibility Code explores a different publishing model: publish for the person while preserving what the machine needs to resolve the knowledge correctly.
From Retrieval to Resolution
Retrieval finds information. Resolution determines what that information means in context.
An answer may require more than locating a relevant passage. A machine may need to determine which entity is being discussed, which facts belong to it, how it relates to other entities, where those assertions came from, when and where they apply, and which information is relevant to the question being answered.
The Visibility Code treats that resolution problem as a publishing problem.
The objective is not to tell machines what to think. It is to expose enough publisher-known structure that unnecessary inference can be reduced.
The Publisher and Consumer Boundary
The Visibility Code draws a deliberate boundary between what publishers control and what machine consumers control.
Publishers control their reference knowledge: identity, facts, definitions, provenance, relationships, applicability, temporal context, derivations, and resolution structure.
Machine consumers control their behavior: what they crawl, index, retrieve, rank, cite, synthesize, transform, or ignore.
This distinction is fundamental. Publishers can improve the representation of their own knowledge. They cannot legitimately claim to control the internal state or behavior of proprietary search engines, language models, answer engines, or agents.
Measure Behavior. Do Not Invent Internals.
The Visibility Code approaches AI visibility as an observable publishing discipline.
Publishing interventions can be recorded. Machine behavior can be sampled. Presence, citation, factual fidelity, identity, attribution, relationships, applicability, provenance, and resolution can be compared over time.
What generally cannot be observed from outside a proprietary system is the internal mechanism that produced those behaviors.
For that reason, The Visibility Code distinguishes observation from correlation, and correlation from causation. The goal is to build an evidence record, not a mythology about how proprietary machines work.
The Public Laboratory
MedicarePlans.com serves as a public production laboratory for this work.
Medicare provides an unusually demanding knowledge environment. Plans, contracts, carriers, benefits, formularies, geographic availability, enrollment, quality measures, eligibility rules, source datasets, and time periods form interconnected information spaces that cannot always be resolved correctly from isolated passages or pages.
Publishing interventions can be deployed to real Medicare information surfaces and followed through search indexing, structured-data recognition, citation behavior, answer fidelity, and multi-turn resolution behavior.
The purpose of the laboratory is not to prove a predetermined theory. It is to observe what happens when publisher-controlled knowledge representation changes—and to record what does not happen as carefully as what does.
WebMEM®
WebMEM is a publisher-side knowledge representation protocol developed from this work.
WebMEM provides a machine-facing layer for preserving publisher-known semantics alongside human-facing web content. Its architecture makes identity, facts, provenance, relationships, applicability, derivations, and other knowledge structures explicit without requiring a second website for machines.
The Visibility Code and WebMEM serve different purposes. The Visibility Code defines the larger publishing discipline and its methods. WebMEM provides one protocol for implementing the machine-facing knowledge layer.
Origins and Stewardship
David W. Bynon created The Visibility Code and WebMEM from work in structured publishing, large-scale Medicare data systems, answer-engine visibility, and publisher-side knowledge representation.
The work evolved through production experimentation rather than from a theory of how proprietary AI systems should behave. Earlier research explored concepts such as machine memory, retrieval conditioning, entity publishing, and structured fragments. Production implementation gradually exposed a more durable boundary: the publisher does not need to control the machine’s memory. The publisher needs to preserve and govern its own knowledge.
That distinction became the foundation of The Visibility Code.
Publish for the person.
Represent for the machine.
Measure the reflection.
Protect the meaning.