Why it matters
Without AI, misinterpreted subsurface signals can drive costly dig decisions, causing service strikes, missed features, or unresolved condition discharges that hold up planning and delivery.
From design to safety
Design
AI guides design and assessment
AI ingests desk-based assessments, site survey outputs, WSI, and grid layouts to flag high-probability features before trenching. It cross-checks against HER records and context coding standards, proposing target horizons and sampling strategies.
- survey grids
- plots
- twin probe arrays
Build
AI supports trench setup and data capture
AI calibrates instruments, defines survey grids, schedules probes, and validates data collection against WSI requirements during trenching. It generates field notebooks, computes Harris matrix entries in real time, and flags sampling gaps for immediate action.
- survey grids
- plots
- resistance meters
Operation
AI analyzes in-field data streams
AI monitors real-time resistance signals, moisture context, and feature markers to guide watching briefs and context sheet annotations. It publishes provisional feature maps for client review and maintains data lineage for post-excavation.
- resistance meters
- twin probe arrays
- soil moisture sensors
Maintenance
AI maintains data integrity and tools
AI flags sensor drift, calibrates devices, updates context records, and ensures archive deposition readiness. It records calibration certificates, firmware versions, and change logs for audit.
- soil moisture sensors
- plots
- wall footings
Safety
AI enhances subsurface safety
AI correlates utility maps, trench stability indicators, and moisture readings to trigger protection measures and planning confirmations. It integrates UXO reports, gas readings, and dewatering status to revise safety briefs.
- wall footings
- survey grids
- plots
Part by part
- 1
resistance meters
- Watch for
- Drifted readings or sudden shifts suggesting contact with buried features.
- AI action
- Detect anomalies and trigger follow-up scan requests and grid refinement.
- 2
twin probe arrays
- Watch for
- Inconsistent probe coupling or impedance changes across probes.
- AI action
- Flag faulty probes, recalibrate, and re-sample affected plots.
- 3
soil moisture sensors
- Watch for
- Abnormal moisture pockets indicating moisture barriers or soil layering.
- AI action
- Annotate moisture gradients and adjust interpretation of resistance data.
- 4
wall footings
- Watch for
- Unmapped foundations or deviations from planned layout.
- AI action
- Cross-reference with plans and trigger focused area trenching on suspect zones.
- 5
survey grids
- Watch for
- Grid gaps or misalignment affecting coverage.
- AI action
- Suggest grid realignment and additional survey points.
- 6
plots
- Watch for
- Missing or inconsistent plotting of resistance anomalies.
- AI action
- Consolidate anomalies into context sheets and update Harris matrix.
Worked example
AI-assisted subsurface mapping for a 4.2 ha development—periods Iron Age to Roman
Illustrative scenario with sample data.
During a risk-averse phase, the client requires rapid, interpretable subsurface maps to inform WSI and trench design while maintaining HER-facing audit trails and sampling rationale.
Signal in
resistanceData: {site: '4.2ha', gridSize: 2.0, probes: 48, date: '2026-03-12', readings: [..]}, moisture: {percent: [12, 24, 18], sensorsOperational: true}, plans: {WSI: 'draft', grid: 'G1'}, context: {HER: 'pending', recorder: 'Field Unit 7'},
System out
AI-generated subsurface map with feature highlights, grid-adjusted trench plan, and a memo for the watching brief, plus a change-log for archive deposition.
Human sign-off: Project manager revises trench routes based on AI maps and forwards the data package to planning archaeologist and HER advisor for formal sign-off.
How to roll it out
Phase 1
Weeks 1 to 2
Baseline data ingest, grid standardization, and initial anomaly scoring, with calibration audits.
Phase 2
Weeks 3 to 6
Live field integration, tracer updates, and provisional feature maps, including risk flags for utilities.
Phase 3
Weeks 7 to 12
Final maps, documentation for WSI, and post-field archiving prep, with archive package automation.
Where people stay in charge
- Decisions on trench design and ground-truthing must be approved by the project manager or planning archaeologist.
- All AI-derived interpretations require human cross-check against site plans and utility records before actions.
- Final feature judgments and archive deposition must be signed off by the curator or client HER liaison before submission.
Questions owners and crews ask
What is the role of AI in earth resistance surveys during design phase?
AI assists data integration from desk-based assessments, grid plans, and WSI to prioritize high-probability targets before trenching, under oversight of the project manager, with traceable decision logs.
How does AI support field data collection and recording?
AI flags data quality issues, recalibrates sensors, and updates context sheets and Harris matrix in real time, ensuring watch briefs are accurate and archiving is consistent.
What safety checks does AI enable during operation?
AI flags overlap with utility maps, monitors trench gradients, and signals when moisture readings indicate potential instability or contamination risk, prompting protective actions.
How is AI integrated into post-excavation analysis and archiving?
AI consolidates resistance maps, feature files, and context sheets into a coherent archive package, ready for post-excavation assessment and HER deposition with complete audit trail.