The Standard
EAC Standard v1.0
EAC Standard v1.0 defines the public methodology for measuring assistant visibility without publishing scoring weights.
The standard exists so the score can be trusted. It documents what AgentFound evaluates, how outcomes are described, and how fixes are verified. It does not disclose the proprietary scoring recipe or signal weights.
Outcome states
| Outcome | Definition |
|---|---|
| Mentioned | The business appears in an assistant response or local result. |
| Cited | The assistant uses a source that supports a claim about the business. |
| Recommended | The assistant includes the business as a suitable option for the user's prompt. |
| Actionable | The answer includes or can reach a next step such as call, book, quote, order, or directions. |
Signal families
The standard evaluates nine signal families:
- AI findability.
- Citation confidence.
- Answerability.
- Authority support.
- Structured data.
- Profile consistency.
- Action path.
- Crawler access.
- Freshness.
Each signal can affect more than one outcome state. For example, crawler access can affect mention and citation, while action path mainly affects recommendation and conversion readiness.
Public scoring principles
- Signals are based on observable assistant behavior and supporting evidence.
- Signal weights are based on observed impact, but weights are not public.
- The same surface can influence multiple signals.
- A fix is not complete until the assistant answer changes.
- Measurement should use prompt families and repeated checks, not a single one-off prompt.
- Methods must distinguish documented platform behavior from research inference.
Verification
Every report item should be traceable to evidence:
- Prompt or query tested.
- Assistant or surface tested.
- Timestamp and location context where relevant.
- Business outcome observed.
- Cited source or evidence object.
- Signal diagnosis.
- Recommended fix.
- Re-scan result.
Gating rule
The standard explains what is measured and why. It does not publish the model weights, scoring thresholds, or full recipe.
Apply the standard through the product
Use the live audit to measure current assistant behavior and prioritize fixes by observed impact.
Run scan