Why it matters
Without AI, misidentified movements lead to incorrect valuations, misdescribed lots, and potential AML or export issues as staff rely on manual notes and memory.
From design to safety
Design
AI-enabled sale planning and structure
The AI reads cataloguing fields, movement categories, dial variants, and historical hammer results to shape the catalog structure, field definitions, and condition rubric, guiding data governance, terms, and the scope of provenance investigations.
- movements
- calibres
- serial numbers
Build
Automated intake and imaging checks
During consignment intake, the system reads serials, calibres, and dial marks from high-resolution photography, cross-checks service histories, and flags missing or inconsistent records; it schedules expert reviews for suspected non-original parts.
- serial numbers
- dials
- service history
Operation
Real-time identification during viewings
In presale viewings and catalog cross-checks, it matches observed features with database templates, flags anomalies, and suggests notes for condition reports; it logs decision rationales for later audit.
- movements
- service history
- replacement parts
Maintenance
Data quality and provenance audits
Routine re-verification of movements, calibres, and serials against updated references; automated provenance notes refresh, export compliance flags re-validated, and batch corrections recorded.
- calibres
- serial numbers
- replacement parts
Safety
Compliance and safeguarding of characteristics
The system ensures AML due diligence, export licensing checks, and provenance verifications are attached to each lot, and logs staff approvals and reviewer identities for sensitive identifications.
- movements
- service history
- replacement parts
Part by part
- 1
movements
- Watch for
- 3D geometry, engravings, and hidden features that could mislead subclass identification or conceal fraudulent alterations
- AI action
- read movement type; classify by calibre family; trigger cross-check workflow; flag discrepancies for human review
- 2
calibres
- Watch for
- unlisted calibres or rare variants missing from reference set; forged or swapped calibres
- AI action
- extract calibre code; map to database; request expert review; propose provenance notes for the lot file
- 3
serial numbers
- Watch for
- faded, altered, or re-stamped numbers risking misattribution
- AI action
- OCR read; validate against master list; flag anomalies for verification; attach image evidence to the lot record
- 4
dials
- Watch for
- dial variants that affect value, period dating, or workshop authentication marks
- AI action
- identify dial style; compare against catalogued examples; generate notes with confidence score and alternative proposals
- 5
service history
- Watch for
- gaps, inconsistent dates, or missing supplier stamps indicating incomplete provenance
- AI action
- aggregate entries; assess completeness; flag missing records for supplier inquiry; request corroborating warranty or service invoices
- 6
replacement parts
- Watch for
- non-original components that alter value or threaten accuracy of condition reports
- AI action
- detect replacements; categorize authenticity; update condition notes; generate impact assessment for reserve guidance
Worked example
Ensure accurate movement and provenance in regional auction
Illustrative scenario with sample data.
A regional house processes a mixed lot of watches and clocks; AI supports identification, provenance, and condition reporting across multiple consignments.
Signal in
movement_type: lever; calibre: 318; serial: SN12345; dial: enamel; service_history: 2015 overhaul; replacement_parts: rotor
System out
identification: movement lever, calibre 318; serial SN12345; dial enamel; provenance noted; service history complete; replacement parts authenticated; condition report updated; reserve guidance issued.
Human sign-off: Head of Cataloguing reviews the AI-generated notes and approves the condition report before publication.
How to roll it out
Phase 1
Phase 1
Baseline data model for movements, calibres, and serials; create reference library and taxonomy.
Phase 2
Phase 2
Pilot in cataloguing benches with two senior cataloguers; integrate with photography workflow and notes.
Phase 3
Phase 3
Full deployment across all consignments; continuous learning from staff reviews and exception handling.
Where people stay in charge
- Decisions about provenance, AML compliance, and export licensing must be signed off by a qualified officer or regulatory body.
- All personal data and sensitive lot details must be handled under approved data governance policies; non-disclosure and audit trails are mandatory.
- Only trained staff may override provenance or compliance flags; every override must be logged with reason and timestamp.
Questions owners and crews ask
What can AI watch to improve accuracy in movement identification?
AI watches movement type, calibre family, serials, dials, and service history to improve accuracy; it flags inconsistencies for expert review and updates condition notes.
Who approves the AI-generated notes before catalog publication?
The head of cataloguing or a senior cataloguer approves AI-generated notes; provenance and AML compliance leaders sign off sensitive identifications.
How does this help with export and regulatory compliance?
The system attaches export licences and provenance checks to lots, logs approvals, and triggers compliance reviews for flagged items.
What happens if the AI finds a mismatch?
AI flags mismatches for human review; the cataloguer reconciles records, updates the lot file, and notifies the compliance officer if needed.