Why it matters
Without AI, misclassification, duplicate items, inconsistent grading, and unclear significance can lead to unnecessary storage, misaligned disposal, degraded provenance trails, and risks to accreditation and insurance coverage during audits.
From design to safety
Design
AI informs store holdings criteria
The system defines data schemas for collection review criteria, duplication flags, condition grade scales, significance benchmarks, disposal record templates, transfer workflow mapping, and audit trails for each decision.
- collection review criteria
- duplicates
- condition grades
- significance assessment
Build
Data integration and workflow setup
Integrations align with the existing catalogue, condition reporting, Bizot lines, and IPM procedures; disposal records and transfer offers are prepared for automated processing with role-based sign-off.
- disposal records
- transfer offers
Operation
Ongoing screening and decision support
Real-time checks compare incoming catalogue updates against predefined criteria, flag duplicates, recalc significance, validate ethics and provenance, and generate disposal packets with recommended actions and reviewer notes.
- collection review criteria
- duplicates
- significance assessment
Maintenance
Review and audit of dispositions
Regular audits verify retention schedules, condition grade baselines, and disposal outcomes; records are updated and archived for provenance, insurance, and future deaccession evaluation.
- disposal records
- transfer offers
Safety
Accountability and compliance monitoring
All decisions require sign-off by the registrar and head of collections; provenance, IPM, Government Indemnity checks, and legal transfer requirements are tracked and auditable.
- significance assessment
- condition grades
Part by part
- 1
collection review criteria
- Watch for
- Missed criteria alignment or incomplete catalogue metadata that could misclassify items across multiple batches.
- AI action
- Reads criteria mappings and flags mismatches; proposes corrective metadata changes and lineage notes.
- 2
duplicates
- Watch for
- Duplicate records with conflicting provenance or storage location causing double retention or misattribution.
- AI action
- Cross-checks item identifiers; flags true duplicates and initiates merge or de-duplication workflow with audit logging.
- 3
condition grades
- Watch for
- Deterioration patterns or inconsistent grading across batches affecting retention decisions.
- AI action
- Analyzes condition reports; suggests grade recalibration, recommended conservation checks, and threshold alerts.
- 4
significance assessment
- Watch for
- Items with uncertain historical value or ambiguous provenance data impacting retention or disposal.
- AI action
- Evaluates significance against catalogue notes; recommends retention, deaccession, or transfer based on policy with justification.
- 5
disposal records
- Watch for
- Incomplete disposal documentation or missing approvals from governance or legal teams.
- AI action
- Generates disposal packets; routes for registrar and governance approval with provenance annotations.
- 6
transfer offers
- Watch for
- Pending transfer proposals without recipient validation or comprehensive legal checks.
- AI action
- Matches offers to holdings; flags missing legal sign-offs and initiates transfer workflow with risk notes.
Worked example
Curation: Rationalising Store Holdings Before Audit
Illustrative scenario with sample data.
A regional museum with a mid-sized collection of 42,000 objects undertakes a systematic rationalisation to reduce redundant storage, clarify provenance, and align with the new accreditation guidelines, ensuring clear significance for retained items.
Signal in
collection_review_criteria={criteria:["provenance","usage history","storage history"],threshold=4}; duplicates_present=true; condition_grades=[{'item':'A123','grade':'Good','date':'2025-06-01'}, {'item':'B457','grade':'Fair','date':'2024-11-12'}]; significance_assessment=[{'item':'A123','score':8},{'item':'C902','score':5}]; disposal_records_missing=false; transfer_offers_pending=true; catalogue_updates=[{'item':'A123','location':'Store Room 3A'},{'item':'C902','location':'Store Room 1B'}]
System out
disposition_summary={items_to_retain=[A123, C902], items_to_dispose=[D560, E312], notes=[provenance clarified, restoration needs recorded, inventory stability verified]}; registrar_approval_needed=true; transfer_offers_generated=true; audit_trail={phase:'design',signoff_by:'Head of Collections on 2026-11-07'}
Human sign-off: Registrar reviews the automatically generated retention list, sign-offs are secured from governance and curatorial leadership, and the accession register is updated with disposal and transfer records, ensuring traceability and compliance.
How to roll it out
Phase 1
Weeks 1 to 2
Data model and data cleansing baseline established; integration with catalogue and condition reporting begins, with initial staffing assignments and training materials issued.
Phase 2
Weeks 3 to 6
Automated duplicate checks and significance scoring enabled; disposal and transfer templates created, plus provisional approval routing and archival naming conventions applied.
Phase 3
Weeks 7 to 10
Full run through with governance sign-off and archival of disposal records, including escalation paths for any contested items.
Where people stay in charge
- All disposal decisions must be signed by the registrar or head of collections with documented rationale and provenance notes.
- Sign-off must reference insurance implications, legal requirements for transfer or disposal, and IPM compliance before execution.
- No item is deaccessioned without a complete provenance record, a retention justification, and an auditable approval trail.
Questions owners and crews ask
What is AI rationalisation in store holdings, and how does it help our store team?
It analyzes collection review criteria, duplicates, condition grades, and significance to guide retention, disposal, and transfer decisions, ensuring accountability, provenance integrity, and compliance with policy.
Who approves disposal decisions when AI flags items for disposal?
Disposal conclusions require sign-off by the registrar and governance or curatorial leadership per policy, with documented rationale and provenance notes.
How does AI handle duplicates and provenance in store holdings?
It cross-checks identifiers, provenance notes, and storage locations, merging true duplicates and surfacing provenance gaps for governance-approved resolution.
What records are produced for accreditation and audits?
Disposal records, transfer offers, and decision logs are created and archived with reference to IPM, Government Indemnity, and retention schedules.