Why it matters
Without AI, misidentification and recording gaps slow finds processing, create archiving discrepancies, and increase risk of missed heritage constraints or misfiled register entries, delaying approvals and archive deposition.
From design to safety
Design
AI supports written scheme and context planning
The system ingests comprehensive context lists, register templates, typology libraries, and chain-of-custody rules to pre-validate recording schemes and trigger watch lists before trenches open. It models potential correlations and crosswalks to HER fields, enabling pre-emptive QA gates, and specifies data capture schemas for each context type, including disaggregated asset strata and stratigraphic linkage to the Harris matrix.
- small finds
- coins
- registers
Build
Real-time identification during trenching
AI reads high-resolution field photography, stereo images from photogrammetry, and portable X-ray scans, links to coins and brooch typologies, and suggests provisional context markers for single-context records. It also logs provenance tags for images, assigns provisional context IDs, and notes handling sequences for fragile items.
- small finds
- metalwork
- X-ray
Operation
On-site processing and QA for finds
The AI cross-references finds with context sheets, uploads provisional identifications, generates descriptive notes, and flags items for curator review and post-excavation assessment, with traceable audit logs. It also creates provisional archive-ready entries and queues items for specialist sign-off.
- coins
- brooches
- registers
Maintenance
Archive readiness and data integrity checks
Continuously validates metadata completeness, image provenance, bin/lot codes, and material preservation notes to support post-excavation assessment, archive deposition, and HER import readiness. It performs scheduled integrity audits of digital seriation keys and updates catalog IDs in the archive management system.
- registers
- X-ray
- coins
Safety
Safe handling and service accountability
AI flags fragile items, monitors handling protocols, and schedules reprocessing needs, ensuring trained personnel verify sensitive finds handling procedures, containment during processing, and compliance with regulatory storage. It also ensures chain-of-custody continuity through all transfer events.
- small finds
- metalwork
- X-ray
Part by part
- 1
small finds
- Watch for
- fragile or corrosion-prone condition changing during handling or imaging; detect micro-fractures during transport or imaging and adjust support
- AI action
- signal condition anomaly and trigger re-packaging workflow, log with batch and handler IDs
- 2
metalwork
- Watch for
- heavy degradation or corrosion patterns altering identification confidence; monitor surface roughness and patina shifts
- AI action
- flag uncertainty level and request expert review with confidence derivative metrics
- 3
coins
- Watch for
- identification ambiguity due to wear or cleaning; track edge wear, mint marks, and planchet condition
- AI action
- propose provisional dating and mint attribution with confidence score and source references
- 4
brooches
- Watch for
- iconography or material mismatches with context; note stylistic period and material provenance
- AI action
- generate provisional typology and flag for specialist sign-off with supporting imagery links
- 5
registers
- Watch for
- missing metadata fields or context links; detect broken provenance chains
- AI action
- mark records as incomplete and assign fill-in tasks to field team with deadlines
- 6
X-ray
- Watch for
- image distortion or inconsistent imaging parameters; verify calibration, focal length, and exposure settings
- AI action
- validate imaging protocol and advise retake or calibration with exposure metadata
Worked example
Site 2.8 ha development finds identification phase
Illustrative scenario with sample data.
Small finds recovery during trenching prompts rapid labeling and provisional typology in a single-context framework.
Signal in
field_id: S2-FT1, item_id: F-013, image: IMG12345.jpg, context: trench-3, typology: unknown; metadata: date, operator, instrument
System out
AI provisional tag: coin, mint: unknown, dating: late Roman; context link established; requires curator sign-off and archive entry.
Human sign-off: Finds supervisor approves provisional tag and routes item to specialist for confirmation and archive scheduling.
How to roll it out
Phase 1
Weeks 1 to 4
Baseline capture of finds data, establishing context linkage, imaging standards, and initial integration with HER and context templates.
Phase 2
Weeks 5 to 12
Automated identification confidence checks, priority queues for specialist review, and integration with post-excavation assessment workflows.
Phase 3
Weeks 13+
Full integration with post-excavation assessment protocols, archive deposition routines, and ongoing curator sign-off for final records.
Where people stay in charge
- Decisions about identification confidence, final typology, and archive deposition must stay with a qualified finds specialist or curator, who has signed-off proficiency.
- All changes to registers and context links require governance sign-off from the project manager or heritage consultant to ensure compliance.
- Sensitive finds and high-risk identifications require additional review by the curator and site safety officer before any action is taken on site or in archive preparation.
Questions owners and crews ask
How does AI assist with small finds before specialist review?
AI provides provisional identifications and context links, highlighting items needing expert sign-off to prevent misclassification and ensure proper archive deposition, with clear justifications and provenance records.
What data standards support this system?
The system uses context sheets, single-context recording, Harris matrix integration, and HER-compatible metadata to maintain traceability and auditability across field and post-excavation stages.
How are images and records stored securely?
Digital records are stored with unique identifiers, access controls, and versioned backups to preserve provenance, support audit trails, and enable secure transfers to archive repositories.
Who approves final records for archive deposition?
The curator or project heritage consultant signs off on final identifications, typologies, and archive deposition packages, with project manager oversight and HER submission confirmation.
Sources and further reading
- CIfA: Standards and GuidanceStandards on finds recording, specialist review, and sign-off procedures for archaeological finds.
- Chartered Institute for ArchaeologistsProfessional practice guidance on accountable records, evidence standards, and handling challenges to archaeological interpretations.
- Archaeology Data ServiceDigital archive deposition requirements, data standards, and OASIS records for excavation archive packages.
- Historic EnglandHistoric Environment Record guidance on data quality, metadata, and archaeological record management.
- HSE: Manual HandlingManual handling risk assessment guidance for lifting, carrying, and repetitive tasks with heavy finds boxes.
- MOLA: ServicesCommercial archaeology services covering finds processing, post-excavation assessment, and specialist reporting workflows.
- US National Park Service: Archeology ProgramFederal archaeology guidance on curation, collection records, and chain-of-custody for archaeological materials.
- GOV.UK: National Planning Policy FrameworkPlanning policy on heritage assets, archaeological interest, and recording requirements for development sites.