# 3Dogs AI Visibility Method

> Make your business easy for AI to discover, understand, trust and act on.

3Dogs treats AI visibility as an evidence path, not a collection of fashionable files. Pointer measures the public experience an AI agent receives. Malamute prepares and tests approved remediation. Re-running the same versioned test can show which named observations changed.

## The five gates

1. **Discover:** Can crawlers reach the site and find the pages, offers and machine-readable resources that matter?
2. **Understand:** Are the category, buyer, product, price, constraints and next step stated plainly in HTML, Markdown and structured data?
3. **Trust:** Do public claims resolve to evidence, current contracts and honest errors rather than soft 404s or unimplemented endpoints?
4. **Act:** Can an agent take a safe next step—scan, ask, compare, prepare a handoff—without inventing authority or exposing a human to an undeclared consequence?
5. **Measure:** Can the same frozen test be rerun after a change, with the raw request and response preserved?

## What is delivered — and what is not yet defined

The public page separates a published entry point from completed delivery. A link can start a scan or briefing; it does not prove that a scan completed, remediation was released, or monitoring is operating.

| Phase | Published role | Evidence required before calling it delivered | Current public proof state | Scope still requiring agreement |
|---|---|---|---|---|
| Baseline | Pointer is the free measurement entry point for a public URL. | Target URL; test and rubric version; run time; named checks; raw request, response or error evidence; resulting observations; and unresolved warnings. | The route and method are published. Fresh runtime completion for a buyer is `NOT_MEASURED` by this local static candidate. | Target ownership is not required for the public scan, but permission, exclusions and handling must be agreed before any private or authenticated source is used. |
| Remediation and verification rerun | Malamute prepares and tests only approved, reversible technical or content changes; release remains under human authority. The same frozen test can compare named observations before and after. | Approved change list; authority and owner; pre-change snapshot; changed files or settings; validation results; rollback; release status; identical target and rubric for the rerun; before/after evidence; remaining warnings. | The method is described. This page proves no production change, deployment, score improvement or completed rerun. | Exact deliverables, acceptance, exclusions, release authority, rollback and price must be written. Work beyond any contracted scope is separately quoted. |
| Monitoring | The Corporate catalog names Pointer monitoring. | Target inventory; cadence; check and rubric version; stale-data rule; failure threshold; alert destination; evidence retention; report format; delivery owner; response expectation; and change-control boundary. | Commercial inclusion is published, but operating cadence, alert delivery, history and service performance are `NOT_MEASURED`; the delivery definition is `UNAVAILABLE`. | The current Terms snapshot does not define these fields. The Corporate “included monthly review” quantity, deliverables, reset or carryover, overage trigger and authorization are also `UNAVAILABLE` and require clarification. |

Any Pointer baseline-improvement service commitment exists only if the current Terms presented at authorization or a signed agreement incorporates the baseline, method, exclusions and remedy. The Terms v1.0 snapshot observed on 2026-08-18 does not incorporate it. Nothing here promises a perfect score, ranking, citation, traffic, revenue, sale or other third-party outcome.

## Technical and content layers

The technical layer covers metadata, canonical URLs, architecture, navigation, internal linking, schema, sitemap and robots policy, accessible semantics, Markdown alternates, API and agent discovery, honest errors, security and authorization boundaries.

The content layer covers category clarity, buyer problems, product hierarchy, answer-focused pages, evidence-linked claims, pricing or qualification rules, comparisons, objections, topical completeness and one unambiguous next step.

## How the products fit

- **Nexus** is the overarching strategic-intelligence product. It runs a researched multi-model debate and preserves dissent.
- **Pointer** is measurement and proof. It makes live requests and prints the evidence.
- **Malamute** is the governed implementation harness for approved, reversible technical and content changes.
- **Bloodhound** is deal intelligence: a ranked, reasoned shortlist against written criteria.

## About scores

A runtime A2A score, a Lighthouse performance score and an AEO/GEO content score are different constructs. 3Dogs preserves each named result with its URL, time, rubric and evidence. A high score in one lane never erases a failure in another.

A public Pointer route for 3dogs.ai is linked below. Current availability, run history and completion require live verification; this local candidate does not prove them.

- Live 3Dogs Pointer result: https://app.3dogs.ai/pointer/scan?url=3dogs.ai
- Run a free scan: https://3dogs.ai/pointer/
- Ask in plain English: https://3dogs.ai/contact/
- Compare plans: https://3dogs.ai/buy/

## Straight answers

### What is the 3Dogs approach to AI visibility?

Measure whether an AI can discover, understand, trust and act on the public offer; show the raw evidence; fix only approved gaps; then rerun the same test.

### How is AI visibility different from SEO?

SEO helps a page appear in search results. AEO and GEO help answer engines understand and cite it. Agent readiness adds whether an AI can take a safe, declared next step.

### Does an llms.txt file guarantee visibility?

No. It can provide a useful content map, but no single file proves that an answer engine will cite a site or that an agent can complete a task.

### Can 3Dogs guarantee a ranking or citation?

No. A measurable Pointer-baseline service commitment applies only when it is incorporated into the applicable Terms or a signed agreement with its baseline, method, exclusions, and remedy. The public Terms version 1.0 observed on 18 August 2026 does not incorporate that commitment. No page promises a perfect score, ranking, citation, traffic, revenue, purchase, or other third-party outcome.
