Why it matters
Without it, misidentification, unverified buyers, AML breaches, and export controls risk legal action, seized consignments, and halted sales due to noncompliant identities and insufficient due diligence.
From design to safety
Design
AI in sale planning compliance
The system defines sanctions, PEP, and adverse-media policy rules, sets screening thresholds, aligns with export controls, and specifies data capture requirements for future consignments, buyer screening, and seller vetting, including data lineage, retention periods, and privacy mappings.
- sanctions lists
- politically exposed persons
- adverse media
- records
Build
Rule ingestion and data readiness
During consignment intake, the platform ingests identity data, ownership documents, passport data, company registrations, public records, and watchlists to tag risk and configure screening workflows, plus schema checks and data quality flags.
- sanctions lists
- politically exposed persons
- adverse media
- records
Operation
Continuous screening in bidding and settlement
As buyers register and bids are logged, automated checks against updated lists run in real time; alerts appear for matches, elevated risk, or need for manual review, and case notes update provenance.
- screening results
- escalation
- records
Maintenance
List management and model revalidation
Regular feed updates, revalidation of risk scoring models, and media feed tuning ensure current standing and trigger re-evaluation of active consignments, approvals, and payments, with versioned policies and rollback plans.
- sanctions lists
- adverse media
- records
Safety
Governance, audit, and export checks
Escalations are documented, approvals signed by the AML officer, and records retained for regulatory review and export-permission checks, with audit-ready reports and tamper-evident logging.
- escalation
- records
Part by part
- 1
sanctions lists
- Watch for
- Mismatched identity data versus entries; name variants not matching; repeated near-miss flags indicating data quality gaps.
- AI action
- Flag potential match for human review and log for audit trail.
- 2
politically exposed persons
- Watch for
- High-risk alias use; residency changes; new relationships with sanctioned jurisdictions; unusual transaction patterns associated with PEPs.
- AI action
- Trigger escalation workflow and halt high-risk activity pending compliance sign-off.
- 3
adverse media
- Watch for
- Negative press linked to buyer or seller; inconsistent attribution; name similarity with ongoing investigations.
- AI action
- Annotate profiles, alert compliance, and request verification before sale.
- 4
screening results
- Watch for
- Low-confidence matches; conflicting data across sources; outdated screening status.
- AI action
- Add confidence score, route to compliance for decision, and update record history.
- 5
escalation
- Watch for
- Repeated false positives; unresolved high-risk flags; urgent export-control concerns.
- AI action
- Generate escalation ticket to compliance officer with recommended actions.
- 6
records
- Watch for
- Incomplete audit trails; missing consent or export licenses; gaps in provenance documentation.
- AI action
- Lock records, append lineage, and prepare export-license package if required.
Worked example
Escalation of a suspected PEP flagged lot
Illustrative scenario with sample data.
A regional gallery consigns a 19th-century painting with a buyer from a high-risk jurisdiction. The system flags a potential PEP connection through adverse media.
Signal in
buyer_id: 78542, name: John R. Doe, passport: X1234567, list_status: suspect, sources: [watchlist_v1, media_v2], risk_score: 0.72, lot_id: 5098, lot_title: 'Portrait of a Gentleman'
System out
Alert generated to AML officer; risk_score 0.72; escalation created; hold placed on lot 5098; export check initiated.
Human sign-off: AML officer reviews, then approves release with enhanced due diligence or places a hold until resolved.
How to roll it out
Phase 1
Weeks 1 to 4: Baseline
Establish data feeds, screening rules, and initial risk scoring with governance without live sales.
Phase 2
Weeks 5 to 8: Pilot
Run parallel to existing processes; refine thresholds and escalation templates based on observed false positives.
Phase 3
Weeks 9 to 12: Production
Full operational adoption; continuous monitoring, audit-ready logs, and routine list updates.
Where people stay in charge
- Blocking or releasing a bidder or consignor must be decided by the AML/compliance officer with full data access.
- All list data, adverse media feeds, and decision logs must be auditable with time-stamped records.
- Escalations require signed off approvals before any sale proceeds, preserving governance and traceability across all levels.
Questions owners and crews ask
Who is responsible for approving flagged buyers or consignors?
The AML/compliance officer signs off on flagged cases after reviewing data, risk scores, and available provenance.
What records must be retained for audits under sanctions rules?
Full screening results, list source data, risk scores, escalation notes, and decision logs must be retained.
How does adverse media affect sale decisions?
Adverse media triggers risk tagging and potential escalation; may lead to temporary holds or further due diligence before proceeding.
What should a saleroom do when a screening result is inconclusive?
Document the uncertainty, request additional verification, and route to compliance for a mediated decision.