Skip to content

AI Provenance

Two years from now, somebody will ask how a particular mapping, control or observation came to exist. Provenance is what makes that answerable.


What is recorded

Every AI interaction produces a record containing:

ElementPurpose
ActionThe specific feature that made the call
Task familyThe coarse grouping the interaction belongs to
ProviderWhich AI provider was used
ModelWhich model
PromptThe exact prompt sent
ResponseThe raw response received
SubjectWhat the call was about — the requirement, control, incident or evidence item
RunWhich run or batch it belonged to
ConfidenceWhere the capability produces one
TokensPrompt and completion token counts
LatencyHow long the call took
StatusWhether it succeeded, and any error
TenantWhich tenant the interaction belongs to

Stamped at write time, not back-filled

Provenance is part of the immutable record, not metadata added later

The subject and run context are stamped onto the interaction at the moment it is written.

Back-filling provenance afterwards would mean the lineage is a reconstruction, which is exactly what it is supposed to replace. A record whose provenance was added later cannot be relied on to say what it claims.

The mechanism is deliberate: the calling feature declares what it is working on, and the telemetry layer captures whatever is declared as the interaction is recorded. An undeclared context simply produces empty columns — it never produces a guess.


Task families

Interactions are grouped into normalised families, which is the level at which the record is actually useful:

FamilyCovers
requirement_extraction, obligation_extractionRegulatory extraction
control_extractionControl source ingestion
mappingCrosswalk proposal
design_adequacyAdequacy checks
evidence_assessmentEvidence-related analysis
compliance_determinationCompliance-related drafting
assurance_rationaleRequirement assurance rationale
risk_determination, finding_determinationRisk and finding analysis
policy_analysisPolicy-related analysis
exception_draftingException justification drafting
audit_assistAudit assistance
inspection_assist, inspection_extractionInspection drafting and extraction
bcm_assist, bcm_scenario_generationResilience assistance
incident_assist, incident_classification, root_cause_analysisIncident assistance
otherAnything outside the above

The action field keeps the raw calling feature; the family is the grouping you would organise an analysis or a review around.


What provenance answers

QuestionAnswered by
How did this mapping come to exist?Mapping source ai_suggested, plus provider, model and prompt version on the record
Which model produced our extracted controls?The control_extraction interactions for that document
A model version turned out to be poor — what did it produce?Every interaction recording that model
Did a person review this AI output?The record's approval history, alongside its provenance
How much AI usage does this tenant have?Aggregate token and interaction counts
Did an AI call fail silently?Status and error on the interaction

Provenance on the artefacts themselves

Beyond the interaction log, the artefacts carry their own provenance fields.

ArtefactCarries
Crosswalk mappingMapping source, AI provider, model, prompt version, AI confidence, rationale
Extracted requirementSource document, source location, extracted text, confidence, provenance metadata
Extracted controlSource lineage, confidence band recorded on the candidate
Inspection candidateSource document, location, extracted text, original wording, confidence
Indicator resultEngine provenance — deterministic, no model involved

Note the last row

Indicator results carry provenance too, recording the engine and version that computed them. That provenance names no model, because no model was involved. Deterministic computation is provenanced for the same reason AI output is: so it can be reproduced.


Confidence, and what it is worth

Where a capability produces a confidence value, it is recorded. It is worth being precise about what it means.

Confidence expresses how strongly the model matched a pattern. It is not a probability that the output is correct, and it is not evidence.

Use confidence forDo not use it for
Triaging a review queueDeciding what needs no review
Routing low-confidence items for closer checkingBatch approval above a threshold
Identifying where source text was unclearJustifying an approval decision

Approving everything above a confidence score reproduces exactly the problem maker-checker exists to prevent.


Tenant isolation

AI interactions are tenant-scoped. Provenance records, prompts and responses belong to the tenant that generated them and are never visible across tenants.

Tenant data is not used to train models. Where a tenant explicitly consents, its interactions may form part of a consented dataset — that is a deliberate, opt-in arrangement, not a default.

See Tenant Data Boundaries and AI Provider Security.


Permissions

ActionPermission
View AI-produced records and their provenance fieldsRead access to the record
Read the audit trailaudit.read
View tenant AI configurationsettings.read

Platform-level AI activity monitoring is a platform administration capability and is not part of the tenant workspace.


Example

Mapping MAP-2026-0233 — control CTL-2026-0041 mapped to obligation A.8.5 as equivalent.

ElementValue
Mapping sourceai_suggested
AI provider and modelRecorded on the mapping
Prompt versionRecorded
AI confidenceHigh
AI rationale"Control requires MFA for remote and administrative access to systems classified critical, matching the obligation's scope and mechanism"
ProposedBy the Compliance Analyst who ran the pass
ApprovedBy the Compliance Manager, three days later
Approval note"Confirmed against the control text. Scope of 'critical' in the control matches the obligation's definition."

Eighteen months later, an auditor asks how the bank concluded this control satisfies this obligation.

The answer is a record: an AI pass proposed it with a stated rationale, using a named model at a named prompt version; a named analyst ran the pass; a different named manager verified the rationale against the control text and approved it on a specific date, with their own note.

Nobody has to remember anything.


OrviQ Enterprise Governance, Risk & Compliance Platform