The engine
AIOTruth is the evaluation and verification engine behind AIOInsights.
It measures the observable signals a business publishes, shows the evidence behind each finding, and refuses to score what it cannot observe. It is deterministic rules, not a generative model.
Engine v2.4.0. Deterministic scoring. Evidence shown.
Can AI systems access, understand, verify, and represent this business from the evidence it publishes?
What the engine does
AIOTruth is the analytical layer that produces the answer AIOInsights shows a business. AIOInsights is the platform and the form experience. AIOTruth is the evaluation system and the methodology behind it. A customer runs a check on AIOInsights and sees the result labeled Powered by AIOTruth.
Where AIOTruth sits
| Property | Role |
|---|---|
| AIOInsights | The public platform and the check experience. |
| AIOTruth | The evaluation and verification engine, and its methodology and evidence. |
| AIOStory | Compares AIOTruth outputs across businesses and locations. |
| AIOFacts | Defines the terms a finding refers to. |
| AIOWeather | Interprets change in the environment the engine measures against. |
| Digilu | Implements and manages the work a finding calls for. |
What it evaluates
AIOTruth runs a set of evaluation types, each behind an AIOInsights check. A new machine standard becomes a new evaluation type, not a new brand.
- Trust Visibility (Live). Evaluates the observable signals a business publishes across clarity, entity consistency, authority, AI discoverability, trust, and local presence.
- Content Credibility (Beta). Evaluates whether a page is readable, attributable, sourced, specific, original, and honestly maintained. It does not claim to prove whether a human or an AI wrote the page.
- WebMCP Readiness (Beta, Early Preview). Inspects the current evidence that a website exposes structured tools for browser agents. Absence is not a current failure, and presence is not a security certification.
Two scores that must not be confused
The engine produces an evaluation result for a specific business, such as the Trust Visibility Score. That is a different measurement from the Research Value Score, which ranks developments in the AIOTruth Research record. One measures a business. The other ranks a piece of news.
Versioning
The engine is versioned. Every result is stamped with the engine version that produced it, so a result stays reproducible even after the engine improves. The current version is Engine v2.4.0. See the changelog.
How to read a finding
Read the methodology for the full account of inputs, retrieval, signals, weighting, failure states, and the limits of each type.