The stronger the claim, the stronger the evidence path must be.

Six vectors, one published weighting.

The Decision Health Index is a weighted function of these six vectors, not a score assigned by judgement. The weighting is published so the index can be recomputed independently. Move any vector to see how.

Observed values are from the demonstration environment (Ironvale Supply). The weights are the published GEON-2.4 weights.

Where the observations come from.

1. 01
   Public AI responses

   Answers observed from six engines against a fixed question set, recorded with engine, date and full response.

2. 02
   Third-party search and market data

   Keyword volume, organic position, competition and cost per click.

3. 03
   Connected first-party analytics

   Search Console and Analytics, where the customer connects them. Improves confidence; never required.

4. 04
   Public competitor information

   Published positions, citations and the sources engines reach for when recommending a competitor.

5. 05
   Public content, reputation and trust signals

   Independent validation, review corpora and cross-source consistency.

6. 06
   Customer-configured business assumptions

   Average deal value, conversion rates and revenue split by decision stage. Always labelled as customer-provided.

7. 07
   Historical GeoRepute scans and action outcomes

   Prior observations and whether an executed action moved the signal it targeted.

Every conclusion carries its own confidence.

Confidence is attached to the individual conclusion, not to the product. Two findings in the same readout can carry different confidence, and they frequently do.

• ●High confidence

  Multiple current and reliable evidence sources support the conclusion.

• ◐Medium confidence

  Evidence is meaningful but partially estimated or incomplete.

• ◔Directional

  Useful for prioritization, not an audited or verified fact.

• ○Insufficient history

  The platform must withhold predictive conclusions until enough evidence exists.

Directional, ranged, and never described as confirmed.

What this system cannot tell you.

1. 1

   AI engine answers vary between runs and between users. Observations are point-in-time samples, not guarantees of what any individual buyer sees.

2. 2

   Commercial exposure is a directional model built on customer-configured assumptions. It is useful for prioritisation and is not an audited financial statement.

3. 3

   Attribution between an executed action and a signal movement is correlational. The system records both and reports the relationship; it does not claim causation.

4. 4

   Competitor authority counts reflect sources engines cite publicly. A competitor may hold private advantages this system cannot observe.

5. 5

   Where history is insufficient, predictive conclusions are withheld rather than estimated.