AI operations & data tooling
Evidence for that status: Retrieval, evaluation and decision-logging components exist inside group applications; they are not yet packaged as a licensable product.
What it does
Retrieval
Corpus management with source versioning, staleness signals and citation back to the document that was actually read.
Evaluation
Test suites that run before a pilot and again on every model change, with regressions reported against a named owner.
Decision logging
The inputs, the policy version, the output and the human action — recorded so a decision can be reconstructed two years later.
White-labelled, and not
Yours
- Brand and the internal-facing interface
- Corpus, policy and evaluation configuration
- Retention configuration to your policy
Not white-labelled
- Model provider relationships and their terms
- Your lawful basis for processing
- The source code, unless a source licence is agreed separately
Integration surface
REST for retrieval and decision submission; a harness runner for CI; an export of the decision record designed for an auditor rather than an analyst.
Who operates what
The licensee owns the corpus, the policy and the lawful basis. We own the software, its upgrades and its defects. Model providers are contracted by the licensee unless otherwise agreed in writing.
- Deployment
- Your cloud · ours · hybrid
- Models
- Provider-agnostic; contracted by you
- Licensing
- Annual + seats
- Support
- Tiered; SLA in contract
Publication gate: capability claims on this page are verified against the running software before publication. Lineage: Built in-house.
Reference deployment
Held back pending written client permission. No named licensee is published without it.
Request a technical evaluation
Engineers on both sides, against your requirements. If the software does not do what you need, this is where you find out.