What is UAMS AI Wayfinding?

Audience: UAMS Communications staff, web stewards, content owners, and department liaisons.

Purpose: Explain what AI Wayfinding is, why UAMS is building it, and what it means for the websites you manage.

The short version

UAMS AI Wayfinding is a standardized set of machine-readable files that UAMS Web Services is publishing across the UAMS web ecosystem. These files help search engines, AI chatbots, and autonomous agents understand what each UAMS domain is, what it contains, and where they should go for the right answer.

It is called Wayfinding because it performs the same function for AI that painted floor lines and color-coded signage perform for a patient walking into a UAMS hospital. The visitor does not know the building. Wayfinding gets them where they need to go.

Why this matters now

People no longer find UAMS content only through Google. They ask ChatGPT where to find a cancer specialist. They ask Claude to summarize clinical trial eligibility. They ask Gemini about UAMS degree programs. When those tools answer, they either pull from authoritative UAMS content or they guess.

Without Wayfinding, an AI agent arriving at uams.edu has to infer where to find provider information, what is a college versus an institute, which news source is current, and which subdomain handles what. That inference is where hallucinations come from.

Wayfinding replaces inference with instruction. On every UAMS pillar that has been migrated to the new file set, the Wayfinding files tell agents, in plain machine-readable language:

• What the domain is
• What it is authoritative for
• Which sibling domains handle related topics
• What an agent should and should not do when citing content from it

What Wayfinding is not

Wayfinding is often confused with the broader UAMS routing architecture, which is a separate and much older part of the ecosystem. A quick distinction:

You do not need to understand the routing architecture to understand Wayfinding. The two layers work together, but they answer different questions.

What it looks like in practice

Pillar domains in the rollout carry three small text files at their root. These files are invisible to regular visitors but immediately useful to any AI agent that knows to look for them.

1. /robots.txt tells crawlers and AI bots what they are allowed to access.
2. /llms.txt is a short, human-readable summary of the domain written for LLMs. Think of it as the handshake.
3. /llms-full.txt is the deep version with detailed directives and links to sibling domains.

If you want to see a deployed example, visit https://communications.uams.edu/llms.txt in your browser. You will see the handshake file that any AI agent would read when it first encounters the Office of Communications and Marketing domain.

What this means for content owners

In most cases, nothing changes about how you create or publish content. Wayfinding is a layer that sits underneath your site and operates automatically.

Three practical notes:

1. The rollout is in progress. Not every pillar and subsite carries the full file set yet.
2. Wayfinding files are governance documents, not regular content. All edits to llms.txt and llms-full.txt are reviewed by UAMS Web Services to maintain cross-domain consistency.
3. Web Services’ stated intention is that new UAMS subdomains should carry the Wayfinding file set before they are considered production-ready. As the rollout matures, this intention is expected to move toward a formally enforced policy in partnership with institutional stakeholders.

Where to go next

• For the operational file-by-file reference, see the Knowledge Base article The Wayfinding File Triad.
• For the full ecosystem map, see the UAMS Digital Ecosystem article.

If you have questions about how Wayfinding applies to a specific property, reach out to UAMS Web Services using the contact form below.