Practical reasons to start
- Enterprise customers asking how data is used and retained
- Retrieval applications with unclear source permissions
- Product changes introducing new vendors or downstream data uses
An example first scope
One customer-facing feature or internal AI use case, with a source register, approved-use boundary and evidence handoff to product and security.
What to bring to an initial conversation
Bring the use-case brief, proposed architecture, intended users and available source or vendor documentation.
Match the work to your requirements
We confirm the domain, systems, jurisdictions, client obligations and internal decision authority before recommending a scope. Regulated or sensitive work may need specialist professionals and a separately agreed delivery team. Do not send confidential data with an initial inquiry.
Relevant services
AI data readiness
Prepare the information your AI will depend on.
Vendor & third-party data review
Understand the data relationship before you sign.
Data inventory & lineage
Know what you have. Know where it goes.
Source permissions that survive a product change.
- Information in scopeApproved documents, product events, support records and the transformations or retrieval indexes used by a selected feature.
- Decision to resolveIs this exact use permitted, for these users and this environment? What happens when a source is corrected or withdrawn?
- Smallest useful pilotChoose one non-production feature. Map authorised sources and users, test contrasting access cases and rehearse source replacement.
- Evidence to keepAn approved-source boundary, version references, access test cases and a list of unresolved reviews for the product owner.
Proposed example only. Actual systems, permissions, delivery capabilities and requirements are confirmed before work starts.