Why it matters
Without AI, planning and execution risk inconsistent stratigraphic reasoning, missed samples, unresolved plan conflicts, and slower documentary trails during post-excavation assessment and reporting.
From design to safety
Design
AI guides strip and sampling plan
The AI reads desk-based assessment data, historic GIS, WSI, and projected trench layouts to validate topsoil stripping zones, feature targets, and sampling density against the written scheme of investigation, while mapping constraints from planning conditions and HER requirements. It cross-checks land registry data, setback distances, and heritage constraints with the project’s risk matrix, and flags any deviations for the design team to resolve before ground deployment, ensuring the trench network aligns with consented mitigation strategies and anticipated feature preservation.
- topsoil strip
- feature mapping plan
- plans
Build
Real-time trench planning support
During trench setup and excavation, the AI compares plans with live machine readings, records context sheets, updates strip maps as layers are exposed, and logs feature exposure milestones for the Harris matrix. It integrates GPS RTK feeds, chainage markers, and local grid corrections, producing a live strip map overlay with color-coded status by context and exposure depth, while maintaining alignment with the written scheme of investigation and contingency allocations.
- topsoil strip
- slot excavation
- plans
Operation
Ongoing mapping and sampling execution
In operation, the AI streamlines context recording, photogrammetry, and sampling log, ensuring single context records align with revised feature maps, updating sampling density, and generating watch lists for mobile teams. It validates context order through Harris matrix sequencing, surfaces inline amendments to sampling strategies for evolving feature visibility, and exports XML/CSV compliant context sheets for the HER submission workflow.
- feature mapping
- sampling strategy
- plans
Maintenance
Plan integrity and archive readiness
Post-excavation, the AI verifies data completeness, context sheets, and archive deposition records against the HER, post-excavation assessment, and CIfA standards for deposition readiness. It audits digital and paper records for legibility, checks box-ticking dependencies for archive accreditation, and schedules subsequent transfers to the archaeology archive with certified courier manifests and tamper-evident packaging logs.
- plans
- sampling strategy
Safety
Safety-focused plan validation
The AI flags risks in trench stability, buried services, and contamination, ensuring planning compliance and sign-off by a qualified person before next works. It simulates trench collapse scenarios, cross-references utility plans, records contamination layers, and aligns safety plans with site welfare provisions and dewatering schedules, triggering immediate reevaluation if thresholds are breached.
- topsoil strip
- slot excavation
- plans
Part by part
- 1
topsoil strip
- Watch for
- soil horizon change, unexpected fill, evidence of earlier features
- AI action
- analyze soil data; trigger context sheet revision and strip plan update
- 2
machine watching
- Watch for
- abnormal machine vibration, excess resistance, hydraulic pressure spike
- AI action
- alert operator, pause machine, and log sensor anomaly
- 3
feature mapping
- Watch for
- new anomaly in geo-data or plan conflicts
- AI action
- re-scan area, update feature map, notify field team
- 4
sampling strategy
- Watch for
- sampling density below plan, missing control samples
- AI action
- adjust sampling plan, issue revised SOP to team
- 5
slot excavation
- Watch for
- slot over-exposure, wall instability, inadvertent feature exposure
- AI action
- trigger trench stop, notify supervisor, replan slot depth
- 6
plans
- Watch for
- plan conflicts or missing context sheets
- AI action
- generate updated strip map plan, distribute to team
Worked example
Coring a buried terrace within a 6.2 ha development trench sequence
Illustrative scenario with sample data.
During an expansion trench sequence, AI-assisted strip and sample planning identifies a buried terrace and optimizes sampling to recover the terrace features before backfilling.
Signal in
siteAreaHa: 6.2, periodsFound: Late Bronze Age, featuresTarget: terrace edge, topsoilDepthcm: 35, samplingDensity: 6m, planVersion: v3.2, equipment: excavator-EX200, GPS: RTK-1
System out
AI-generated update: terrace feature mapped, sampling density increased to 4m around terrace edge, new plan version v3.3 issued, context sheets updated, archive deposition plan adjusted.
Human sign-off: Project manager reviews the updated plan, validates HER alignment, and authorizes field team to implement v3.3 changes.
How to roll it out
Phase 1
Weeks 1 to 4
Baseline data capture, plan validation against WSI and HER requirements.
Phase 2
Weeks 5 to 8
Active trenching with AI-assisted mapping, in-field updates to plans and context sheets.
Phase 3
Weeks 9 to 12
Final sampling optimization, post-excavation data reconciliation, archive deposition prep.
Where people stay in charge
- Decisions about impact to the Written Scheme of Investigation must stay with the project manager or planning archaeologist.
- All AI-generated updates affecting HER compliance require human sign-off and client consultation.
- Safety-critical decisions, including trench reconfiguration or stopping work, must be validated by the site supervisor or responsible licensed archaeologist before implementation.
Questions owners and crews ask
What data does the AI read to plan strip and sampling?
The AI reads the written scheme of investigation, desk-based assessment, historical maps, GIS layers, plan outlines, topsoil thickness data, and live excavation readings to guide strip and sampling decisions.
Who approves AI-generated plan changes?
The project manager, in collaboration with the client’s heritage consultant or planning archaeologist, approves changes that affect HER compliance and post-excavation strategy.
How does AI help with context recording in trenches?
It aligns context sheets with feature maps, photogrammetry outputs, and single context recording conventions, and flags mismatches for quick in-field correction.
How is safety enhanced by AI in trench operations?
AI monitors trench geometry, buried services signals, soil moisture, and gas readings, triggering early stop signals and escalation to the site supervisor when risk thresholds are crossed.
Sources and further reading
- CIfA: Standards and GuidanceStandards and guidance on written schemes, excavation method, and recording expectations for archaeological fieldwork.
- Historic EnglandArchaeological science guidance on soil, deposit, and stratigraphic recording for excavation projects.
- GOV.UK: National Planning Policy FrameworkPlanning policy on heritage assets, archaeological interest, and recording requirements for development sites.
- HSE: ExcavationsTrench collapse hazards, shoring, and protective measures for people working in excavations.
- HSE: ConstructionSite welfare facilities, construction duties, and plant safety for workers on archaeological sites.
- HSE: AsbestosGuidance on hazardous materials found in made ground and contaminated site work.
- Archaeology Data ServiceDigital archive standards, data formats, and deposition guidance for excavation records and OASIS entries.
- Chartered Institute for ArchaeologistsAccreditation requirements, Registered Organisation status, and archive practice guidance for archaeologists.