Shared product contract
Every product must produce the same minimum trust structure.
| Contract layer | Required behavior | Buyer value |
|---|---|---|
| Uploaded or sample dataset | Every run identifies whether results come from uploaded data or governed sample data. | Prevents demos from being confused with customer-specific proof. |
| Source Fit Score | The raw uploaded file is scored against the selected engine before user correction. | Shows the buyer whether the export is naturally close to the engine requirement. |
| AI Match Score | Column aliases are inferred, scored, and exposed before execution. | Makes automated mapping explainable instead of silently trusting a black-box import. |
| Mapped Readiness Score | After the user confirms or corrects fields, readiness is recalculated before the engine runs. | Prevents poor mapping from producing overconfident diagnostics. |
| Normalization plan | Each mapped field shows source column, original sample, normalized preview, and rule. | Lets data stewards verify how the source file will be interpreted. |
| Mapped inputs | Required and optional fields are mapped before analysis. | Shows whether the source export can support the diagnostic. |
| Validation issues | Missing fields, weak completeness, and blockers are visible before scoring. | Prevents false confidence from poor data. |
| Analyzer finding | Each finding has category, severity, confidence, impact, and action. | Turns raw data into a decision object. |
| Evidence record | Each evidence row includes source reference, fields, value, confidence tier, and reason codes. | Lets buyers defend the finding in review. |
| Score formula | Scores expose formula and component inputs. | Prevents random or black-box score claims. |
| Assumptions and limitations | Benchmark, estimate, and uploaded-data result are separated. | Keeps CFO and audit usage safe. |
| Human review | Low-confidence or high-risk findings route to review before action. | Protects against unsafe ERP, procurement, or maintenance changes. |
| Action tracker | Findings become owner, priority, status, due window, and source-finding actions. | Creates accountable work from evidence. |
| Score history | Runs write score snapshots by engine, date, score, risk, and sample/real flag. | Creates recurring improvement visibility. |
| No ERP write-back | Industrial IQ does not autonomously change SAP, Maximo, Oracle, EAM, CMMS, procurement, or inventory systems. | Trust boundary remains explicit. |
