Why it matters
Without this approach, ad hoc concentrations, uneven gel diffusion, or pH drift can cause irreversible damage, solvent misjudgments, or disputed insurance claims and loan conditions.
From design to safety
Design
AI informs protocol conception
AI analyzes historical gel data, enzyme selectivity, substrate sensitivity, and cross-lettering risks to propose concentrations, pH targets, and cleaning times for reversible treatment options, with substrate-compatibility notes for varnish or consolidant layers.
- enzyme solution reservoir
- agarose gel bath
- gellan gum gel sheet
Build
Prototype protocol validation
The team configures the protocol in a controlled bench setup, records initial readings, and adjusts parameters before bench-scale trials, including trial runs on mock substrates with standardized colorants.
- pH probe
- conductivity sensor
- residue clearance station
Operation
Live protocol execution
During treatment, AI monitors real-time pH, conductivity, enzyme activity, and gel diffusion, updating the treatment log and triggering alarms if limits breach, while maintaining optional video-logging of substrate response.
- enzyme solution reservoir
- agarose gel bath
- pH probe
Maintenance
Calibration and stock checks
Regular calibration of sensors, verification of gel batches, and revalidation of concentrations ensure reproducibility and traceability, with batch records linking to PACR-compliant treatment reports.
- conductivity sensor
- residue clearance station
- gel preparation station
Safety
Ethical and regulatory controls
The system enforces reversibility, tracks COSHH compliance, and stores consented treatment proposals signed by the accredited conservator, with auditable change control and access permissions.
- pH probe
- enzyme solution reservoir
- gel preparation station
Part by part
- 1
enzyme solution reservoir
- Watch for
- Declining enzyme activity, contamination, or precipitation altering cleaning efficacy.
- AI action
- AI flags low activity or contamination and suggests batch discard or retitration.
- 2
agarose gel bath
- Watch for
- Gel diffusion inconsistencies or drying edges affecting cleaning uniformity.
- AI action
- AI detects diffusion anomalies and recalculates immersion time and gel thickness.
- 3
gellan gum gel sheet
- Watch for
- Gel fracture, shrinkage, or improper integration with substrate.
- AI action
- AI adjusts gel formulation and contact plan to maintain contact and reversibility.
- 4
pH probe
- Watch for
- Drift above acceptable range, sensor fouling, or response lag.
- AI action
- AI initiates calibration flag and updates target pH in treatment log.
- 5
conductivity sensor
- Watch for
- Unstable conductivity indicating incomplete solvent exchange or residue.
- AI action
- AI triggers extended rinse plan and logs conductivity trajectory.
- 6
residue clearance station
- Watch for
- Accumulation of residues on substrate surface or gel residue adhesion.
- AI action
- AI prompts additional rinse cycles and records residue metrics.
Worked example
Enzyme and Gel Cleaning Protocol for a fragile panel
Illustrative scenario with sample data.
A fragile wood panel with former varnish layers requires gentle cleaning using enzyme and gel methods, with risk of delamination and color change.
Signal in
substrate=panel; varnish=protein-based; enzyme=protease A; gel=agarose; pH=6.5; conductivity=12 μS/cm; residue=moderate; dyeTest=blue
System out
protocol approved; cleaning plan logged; gel diffusion and enzyme activity thresholds set; disposal and disposal log created; color-change risk notes added.
Human sign-off: Head of Conservation reviews protocol design, approves treatment proposal, and signs off the corresponding condition report.
How to roll it out
Phase 1
Weeks 1 to 4: Baseline
Establish baseline data, calibrate sensors, validate gel batches, and document substrate variants against control panels to ensure repeatable results.
Phase 2
Weeks 5 to 8: Bench to site
Scale up to pilot treatments on non-collection objects; refine AI alerts, logging workflows, and traceability to condition catalogs.
Phase 3
Weeks 9 to 12: Full integration
Incorporate AI protocol templates into standard treatment proposals; align with COSHH, PACR, and ethical consent procedures across projects.
Where people stay in charge
- Decisions about concentration, exposure time, and solvent selection must be made by a qualified conservator or conservation scientist.
- All handling of enzymes and gels must comply with COSHH and institutional safety policies, with final approvals by the accredited conservator.
- AI outputs are advisory; human sign-off is required for any treatment proposal or protocol change.
Questions owners and crews ask
What data does the AI use to design enzyme cleaning protocols?
The AI analyzes historical treatments, enzyme activity readings, gel diffusion data, pH and conductivity measurements, residue metrics, and substrate-specific risks to propose concentration ranges and contact times, with traceable source notes.
How is reversibility ensured in AI-guided cleaning?
Protocols are designed to allow undoing steps, with clear records of materials used and withdrawal plans, and conservative thresholds monitored by the conservator, plus explicit consent logging for each step.
Who signs off on AI-generated treatment plans?
The accredited conservator or the head of conservation signs off, ensuring compliance with PACR accreditation, ethical standards, and curator-approved project briefs.
What safety data are recorded with AI-guided methods?
COSHH compliance, material safety data, exposure risks, PPE requirements, and substrate sensitivity notes are logged and linked to the treatment proposal for audits.