← Trust Centre

AI governance

Evidence people can review. Decisions people still own.

Your team needs to know where an AI answer came from and who should check it. Our public examples use sample data. Evidence access, AI providers and permitted uses must be agreed for an evaluation.

Public information · Reviewed 27 September 2026

For each AI task, agree its purpose, allowed sources and data, provider terms, testing, monitoring and human review. AI cannot make clinical, regulatory, vendor selection or commercial decisions.

Check which documents are available →

The public examples use sample data. Check what is available for an evaluation and which security and quality documents can be shared before sending sensitive information.

Evidence to review

What your team needs to know.

Use this list to plan your checks. A listed safeguard is something to verify, not confirmation that it is in place or that supporting documents can be shared.

Review areaCurrent statusPublic information
01Intended useIntended use

Finding and comparing evidence, and exploring possible effects, can support study design decisions. Clinical and commercial approvals remain with authorised people.

02Source and review historyReview requirements

Check how an AI answer was produced: the model, sources, dates, relevance, assumptions, limits and human review needed. Where approval applies, verify that the record identifies the approving person.

03Matching relevant studiesReview requirements

Assess study comparisons by indication, phase, population and design. Examine source coverage and the basis of each comparison for your use case.

04Human oversightReview requirements

Authorised people review the reasons for a change and approve the exact protocol version.

05Provider and data useTo be agreed

Model providers, retention, training restrictions and data routes must be agreed, documented and verified before production. This page does not confirm one setting for all services.

06Evaluation and monitoringEvaluation required

Testing for each use case, its limitations, review thresholds and change monitoring are defined for the intended production scope.

07Change controlDocuments not yet confirmed

Significant changes to models, AI instructions, providers or rules need a recorded review that reflects their possible impact.

Common questions

Answers for your supplier review.

These answers describe what is publicly confirmed. Check the agreed safeguards and supporting evidence for the service you plan to use.

Request supporting documents

No. Tenvia presents decision support. Clinical, regulatory, quality and commercial authority remains with appropriately authorised people.

Documents to request

Ask what evidence is available for your review.

The list below shows documents you may need for a supplier review. We must confirm which are available and can be shared before granting access.

01AI system overviewCheck availabilityOn request

02Evidence source and retrieval modelCheck availabilityOn request

03Approach to matching relevant studiesCheck availabilityOn request

04Intended use and limitations statementCheck availabilityOn request

05Provider and data use scheduleCheck availabilityOn request

06Evaluation summaryCheck availabilityOn request

07Human oversight and change control modelCheck availabilityOn request

Security & supplier review

A clear path
to your review.

Explore the product, discuss a study or request assurance evidence. Choose the conversation your team needs.

What would you like to review?

01 / Choose your review
Confidentiality agreement and approved access

Tell us which security, privacy, quality or architecture evidence your team needs to review.

Your review

02 / Describe your priorities

Please keep patient information and confidential study documents out of this request.

Submit your enquiry directly through HubSpot.
Tenvia will contact you to follow up.

Prefer a direct conversation?

Contact Tenvia
Continue your reviewClinical & quality
Open full guide