Catalog health score
Use the score as a diagnostic signal that needs evidence review.
Industrial IQ | AI2COE | PartsCleanse AI
PartsCleanse AI Catalog Diagnostic Report
Sample report output map
How to evaluate the Catalog Intelligence output bundle
This visual map helps buyers trace the sample report as a product output: source file, field mapping, evidence rows, confidence context, report pack, and governed owner action stay connected before any private upload.
duplicate family, matching reason, discriminator conflict
Inspect source context, confidence, assumptions, and limits.
PartsCleanse AI Catalog Diagnostic Report
Use HTML and PDF views for buyer committee review.
Owner, status, next review
Keep remediation buyer-controlled after the sample is replaced with uploaded data.
Baseline and rerun movement
Use recurring reviews to see what changed after accepted actions.
CSV, mapping, data dictionary
Give data owners the field expectations before private upload.
Output boundary: this sample proves output structure and workflow only. Uploaded operational data is required before customer-specific findings can be reviewed.
Diagnostic evidence experience map
PartsCleanse AI shows the buyer the product journey from upload to owner-reviewed action.
The UX is deliberately evidence-first: input data, field mapping, validation, analyzer logic, evidence rows, score, report, and action ownership are visible before any source-system change is considered.
Buyer starts with a read-only operational export, not system credentials.
Required fields, aliases, completeness, and unmapped columns stay visible.
The engine separates usable evidence from missing, weak, or out-of-scope data.
Catalog Intelligence runs the diagnostic logic for catalog health score.
Findings stay tied to source context, reasons, assumptions, and limitations.
Scores are diagnostic signals for review, not official ratings or promised outcomes.
Trust boundary: sample mode demonstrates structure only. Uploaded-data diagnostics remain read-only, source-backed, confidence-tiered, and human-reviewed. No ERP write-back.
Evaluation kit
Download the supporting assets for this Catalog Intelligence sample.
Use the HTML sample to inspect report structure, then use the supporting files to understand what data owners would prepare before replacing sample assumptions with uploaded-data evidence.
PDF
Offline review artifact for buyer committees.
Sample CSVDeterministic sample input used for testing the workflow.
MappingColumn mapping guide for source-fit review.
DictionaryField definitions and expected diagnostic meaning.
Product pageReview the Catalog Intelligence engine page.
Trust boundary: sample assets are deterministic demonstration files. Live diagnostics remain read-only, require owner review before action, and do not write back to ERP/EAM/CMMS systems.
Executive Summary
0.0Catalog health score
Board-Level RiskRisk level
Medium ConfidenceConfidence
32Evidence records
Source mode: Sample dataset result. Sample results demonstrate workflow and report structure; uploaded-data results replace assumptions with customer source evidence.
Score interpretation: Lower health or readiness scores indicate higher unresolved exposure, weaker readiness, or stronger review need. Scores are deterministic and derived from mapped source fields, findings, evidence, and component inputs.
Executive interpretation: Focus on asset integrity, turnaround readiness, working capital exposure, and audit-safe ERP preparation.
Report delivery and governance controls
Email
Report email path
Authenticated Industrial IQ runs attempt branded report email delivery and retain delivery status in the platform report inventory.
Review
Human review required
Low-confidence or high-impact findings should be accepted, rejected, assigned, deferred, or marked needs-more-data before remediation.
ERP safety
No ERP write-back
Industrial IQ produces evidence and recommendations only. It does not autonomously change SAP, Maximo, Oracle, EAM, CMMS, procurement, inventory, or asset records.
Data handling: Uploaded source files are processed to generate the diagnostic report pack and then purged. Summary metrics and audit metadata may be retained for governance.
CFO command view
201906.25Capital exposure signal
24228.75-56533.75Recoverable range
36343.12Annual leakage signal
Board-Level RiskBoard attention band
Diagnostic components
6DUPLICATE GROUPS
100.0DUPLICATE RATE
0MISSING DESCRIPTIONS
0UOM INCONSISTENCY GROUPS
1SUPPLIER ALIAS CLUSTERS
0SPECIFICATION GAP ROWS
0OEM REFERENCE GAP ROWS
154200.0ESTIMATED DUPLICATE VALUE
Score formula: 100 - duplicate_rate*1.2 - missing_descriptions*0.2 - uom_inconsistency_groups*2.5 Random score used: False
| Score input | Value |
|---|---|
| duplicate groups | 6 |
| duplicate rate | 100.0 |
| missing descriptions | 0 |
| uom inconsistency groups | 0 |
| supplier alias clusters | 1 |
| specification gap rows | 0 |
| oem reference gap rows | 0 |
| estimated duplicate value | 154200.0 |
| duplicate prevention ready | True |
| golden record method | description completeness + manufacturer + MPN + value-bearing signal |
| validation issue count | 0 |
| validation blocker count | 0 |
| prompt injection signal count | 0 |
Product maturity and competitive depth
Competitive position: Competes with SPARETECH, Verdantis, Prometheus, SAP MDG, and MDM suites by staying diagnostic-first, export-first, no-write-back, and board-reportable.
| Priority | Implemented product capability |
|---|---|
| P0 | Interchangeability classification: exact duplicate, probable duplicate, same OEM reference, substitute candidate, and unsafe match.; Golden-record candidate generation with completeness, OEM/MPN, specification, value, and UOM reasons.; Supplier alias, canonical UOM, OEM/part-number, taxonomy, and duplicate-prevention evidence. |
| P1 | Taxonomy/specification extraction and missing-attribute backlog by material family.; Reviewer queue by value, confidence, unsafe-match risk, and duplicate-family size.; Cross-site catalog health trend and duplicate-prevention readiness score. |
| P2 | Reference enrichment readiness for OEM catalog, supplier catalog, UNSPSC, eCl@ss, and multilingual descriptions.; Catalog knowledge graph connecting material, OEM, supplier, UOM, site, specification, and review status.; Continuous duplicate-prevention gate for future item creation governance. |
Role packaging
| Package | Engines | Decision supported |
|---|---|---|
| CFO Pack | FinanceMind AI, PartsCleanse AI, InventoryMind AI, ProcureMind AI | Fund value realization only after capital exposure, carrying cost, procurement leakage, and review confidence are visible. |
| CIO / ERP Pack | ReadyMind AI, GovernanceMind AI, PartsCleanse AI, AssetMind AI | Prove ERP, data, governance, and AI readiness before integration, automation, or migration decisions. |
| Procurement Pack | ProcureMind AI, PartsCleanse AI, InventoryMind AI, FinanceMind AI | Convert supplier, PO, duplicate stock, and price-variance evidence into sourcing action without unsupported savings claims. |
| Maintenance / Reliability Pack | ReliabilityMind AI, AssetMind AI, InventoryMind AI, PartsCleanse AI | Prove work-order readiness, asset-to-part coverage, critical-spare availability, and false-stockout risk before maintenance or reliability programs scale. |
Advanced product insights
| Product output | Diagnostic value |
|---|---|
| interchangeability classes | {"exact duplicate": 6} |
| golden record logic | description completeness + manufacturer + MPN + value-bearing signal |
| unsafe match control | Specification, UOM, size/material/pressure tokens, OEM/MPN, and confidence tier keep risky families in human review. |
| reference enrichment readiness | {"duplicate_prevention_ready": true, "oem_reference_gap_rows": 0, "specification_gap_rows": 0} |
Buyer committee views
CFO
Can quantified exposure justify a diagnostic or remediation budget?
PartsCleanse AI shows 201906.25 as the current capital or leakage signal before owner review.
Next question: Which findings have enough confidence and value to enter the financial business case?
COO
Which findings threaten operational continuity, site readiness, or uptime?
6 high-attention findings require operational owner review.
Next question: Which findings must be resolved before the next outage, shutdown, or planning cycle?
CIO
Is the data ready for governed AI without ERP write-back risk?
Industrial IQ produced evidence from exports only and did not change ERP, EAM, CMMS, or procurement systems.
Next question: Which missing fields or governance gaps should be fixed in the next export?
Procurement
Where do supplier, purchase, or stocked-but-purchased signals need review?
Procurement actions should be evidence-led and routed through human review before supplier action.
Next question: Which supplier or purchase findings are defensible enough for category review?
Maintenance
Will spare availability and catalog quality support maintenance execution?
Maintenance should use the evidence queue to protect planned work and critical assets.
Next question: Which findings block planned work, shutdown readiness, or critical equipment coverage?
Board
Is this risk material enough to fund recurring diagnostic intelligence?
The result is diagnostic evidence, not an autonomous system change or unsupported ROI claim.
Next question: Should leadership fund the next diagnostic cycle, review queue, or remediation scope?
Evidence graph
Model: Source Record -> Finding -> Evidence -> Confidence -> Business Impact -> Recommended Action -> Review Status -> Report -> Score History
41 nodes | 40 evidence relationships. This graph links uploaded source rows to findings, confidence, business impact, recommended actions, report output, and score history.
Renewal value view
201906.25EXPOSURE IDENTIFIED
7REVIEW QUEUE SIZE
6ACTIONS CREATED
0ACTIONS REVIEWED
16152.5CONSERVATIVE VALUE REALIZATION
32305.0BASE VALUE REALIZATION
monthly for high-risk sites; quarterly for controlled sitesNEXT REVIEW CADENCE
Recurring value interpretation: Compare this run against the next upload to show exposure reviewed, actions completed, score movement, and remaining risk.
Key Findings
Trust control: Each finding must be interpreted with its confidence, evidence count, mapped fields, and source records. Similar-looking industrial records may still require owner review before action.
| Analyzer | Finding | Severity | Confidence | Evidence | Action |
|---|---|---|---|---|---|
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Duplicate Family Detector | 4 likely duplicate catalog records in one family (exact duplicate) | HIGH | 84% | 4 | Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records. |
| Supplier Alias Analyzer | 1 supplier alias clusters need normalization | MEDIUM | 74% | 2 | Normalize supplier aliases and map preferred supplier identity before procurement or catalog remediation. |
Evidence Records
| ID | Confidence tier | Severity | Description | Value | Source | Reason codes |
|---|---|---|---|---|---|---|
| E-d34f3533 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5850.0 | row:1:MAT-001-001 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-7efc7ff2 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5850.0 | row:7:MAT-001-007 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-29067833 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5850.0 | row:13:MAT-001-013 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-1fac4b80 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5850.0 | row:19:MAT-001-019 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-e0cc9d99 | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 5850.0 | row:7:MAT-001-007 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-38e0186c | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5500.0 | row:2:MAT-002-002 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-3ee8d3de | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5500.0 | row:8:MAT-002-008 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-3aa1bbfe | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5500.0 | row:14:MAT-002-014 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-607e143b | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 5500.0 | row:20:MAT-002-020 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-5b01f3c5 | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 5500.0 | row:14:MAT-002-014 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-a55122ad | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 7350.0 | row:3:MAT-003-003 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-8271130c | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 7350.0 | row:9:MAT-003-009 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-615b89d6 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 7350.0 | row:15:MAT-003-015 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-409ab824 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 7350.0 | row:21:MAT-003-021 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-b23709ee | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 7350.0 | row:15:MAT-003-015 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-5d745ed2 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 6750.0 | row:4:MAT-004-004 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-c15118cf | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 6750.0 | row:10:MAT-004-010 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-eab52378 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 6750.0 | row:16:MAT-004-016 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-19bafbcf | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 6750.0 | row:22:MAT-004-022 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-c0ac6354 | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 6750.0 | row:22:MAT-004-022 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-84583d0c | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 8850.0 | row:5:MAT-005-005 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-bcd121c4 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 8850.0 | row:11:MAT-005-011 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-95350269 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 8850.0 | row:17:MAT-005-017 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-14714d53 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 8850.0 | row:23:MAT-005-023 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-8e24a14f | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 8850.0 | row:23:MAT-005-023 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-9fa3c490 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 4250.0 | row:6:MAT-000-006 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-e869a8aa | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 4250.0 | row:12:MAT-000-012 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-54338d5f | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 4250.0 | row:18:MAT-000-018 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-8d8022d2 | Medium Confidence | HIGH | 4 likely duplicate catalog records in one family (exact duplicate) | 4250.0 | row:24:MAT-000-024 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-c7eb3550 | High Confidence | MEDIUM | Golden-record candidate selected from completeness, OEM/MPN, and value signals | 4250.0 | row:6:MAT-000-006 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-cce97bd5 | Medium Confidence | MEDIUM | 1 supplier alias clusters need normalization | 2925.0 | row:1:MAT-001-001 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
| E-d530eeb6 | Medium Confidence | MEDIUM | 1 supplier alias clusters need normalization | 4400.0 | row:2:MAT-002-002 | material-id-present, description-signature, manufacturer-signal, oem-mpn-signal, supplier-alias-signal, uom:EA, site-context, value-bearing-row, spec-token-match |
Recommended Actions
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
P0
Review the duplicate family, retain one controlled material record, and retire or cross-reference redundant records.
Owner: CFO | Due: 30 days
Mapping and validation
| Input | Source column | Completeness | Confidence | Reason |
|---|---|---|---|---|
| description | description | 100.0% | 100.0% | Exact ERP alias |
| material_id | material_id | 100.0% | 100.0% | Exact ERP alias |
| manufacturer | manufacturer | 100.0% | 100.0% | Exact ERP alias |
| mpn | mpn | 100.0% | 100.0% | Exact ERP alias |
| supplier | supplier | 100.0% | 100.0% | Exact ERP alias |
| uom | uom | 100.0% | 100.0% | Exact ERP alias |
| quantity | quantity | 100.0% | 100.0% | Exact ERP alias |
| unit_cost | unit_cost | 100.0% | 100.0% | Exact ERP alias |
| site | site | 100.0% | 100.0% | Exact ERP alias |
| currency | currency | 100.0% | 100.0% | Exact ERP alias |
Source fit, AI match, and normalization
100%Source fit score
100.0%AI match score
100%Mapping readiness score
94.3%Diagnostic confidence score
Workbench interpretation: Source fit measures whether the uploaded file contains recognizable inputs. AI match measures column-mapping confidence. Diagnostic readiness measures whether the normalized mapped data can support trustworthy engine output.
| Quality signal | Value |
|---|---|
| source fit score | 100 |
| ai match score | 100.0 |
| diagnostic readiness score | 100 |
| required mapped | 1 |
| required total | 1 |
| optional mapped | 9 |
| optional total | 9 |
| required completeness | 100.0 |
| row count | 24 |
| column count | 30 |
| blockers | 0 |
| warnings | 0 |
| source fit band | Strong |
| ai match band | Strong |
| readiness band | Strong |
| diagnostic confidence score | 94.3 |
| diagnostic confidence band | Strong |
Normalization plan
| Engine field | Source column | Original sample | Normalized preview | Rule |
|---|---|---|---|---|
| Description | description | Oil & Gas pump bearing seal kit model 1 stainless 4 inch | OIL & GAS PUMP BEARING SEAL KIT MODEL 1 STAINLESS 4 INCH | Normalize case, abbreviations, punctuation, industrial units, specification tokens, and obvious spacing noise. |
| Material Id | material_id | MAT-001-001 | MAT-001-001 | Trim whitespace, preserve leading zeroes, normalize item/material identifiers, and keep original source reference. |
| Manufacturer | manufacturer | Flowserve | FLOWSERVE | Normalize manufacturer/OEM aliases and common spelling variants while preserving source value. |
| Mpn | mpn | FS-0001 | FS-0001 | Normalize part-number separators, case, model tokens, and OEM reference variants. |
| Supplier | supplier | Industrial Supply Co | INDUSTRIAL SUPPLY CO | Normalize supplier aliases and vendor naming variants for leakage and overlap analysis. |
| Uom | uom | EA | EA | Map EA, Each, PCS, SET, M, MM, IN, Pack, and similar variants into canonical UOM families. |
| Quantity | quantity | 3 | 3 | Parse numeric quantity, keep negatives for audit context, and separate blank/zero from missing. |
| Unit Cost | unit_cost | 975 | 975 | Parse unit cost, retain source currency, and separate uploaded value from benchmark assumption. |
| Site | site | Plant-2 | Plant-2 | Normalize plant, site, storeroom, facility, depot, or operating-unit labels. |
| Currency | currency | USD | USD | Normalize ISO currency code or use country/system assumption with visible limitation. |
Assumptions and limitations
Assumptions
- Uploaded data is treated as the source of truth for this diagnostic run.
- No ERP write-back is performed. Outputs are recommendations and evidence records only.
- Financial estimates use uploaded values where available and conservative assumptions otherwise.
- Industry language is adjusted for Oil & Gas: plants, wells, refineries, shutdowns, turnarounds, and asset integrity.
- Workbench scores were calculated before and after engine execution: source fit 100%, AI match 100.0%, mapping readiness 100%, diagnostic confidence 94.3%.
- Public sample report: deterministic AI2COE sample data was used. Replace with uploaded customer data for customer-specific findings.
Limitations
- Results are diagnostic signals, not final accounting entries.
- Low-confidence findings require human review before remediation.
- Missing source fields reduce confidence and may suppress some analyzers.
- Benchmarks are labelled assumptions unless validated by uploaded data.