Available now
Public buyer answers
What the website offers, what data you can share and what to confirm before customer use.
Read the answers →Tenvia Trust Centre
Intelligence informs.
People decide.
Tenvia connects protocol design, evidence, team review and decision history. Accountable specialists make the study decisions. Explore the security, data use and commercial arrangements for your team.
Direct answers
Public facts.
Clear outstanding questions.
- Can I upload a confidential protocol here?
- No public protocol-upload service is offered. Use the contact form for business enquiries only. Agree confidentiality, data handling and access before sharing study files.
- Which AI models process the public demo?
- The guided RD-204 simulation runs in your browser with prepared content and no live AI. This does not establish the models or providers for a future customer service.
- Will my customer protocol be used for model training?
- This public pack does not confirm whether customer data is used for AI training or how long it is kept. Obtain written terms covering Tenvia and each AI provider before sharing study material.
- Are SSO and MFA supported for customer access?
- This pack does not confirm customer sign-in options. Ask for evidence covering single sign-on, extra sign-in checks, session controls and removal of access for your product version.
- Does Tenvia sell personal data?
- This pack does not provide a confirmed answer for the customer service. The website privacy notice describes business enquiry handling; the allowed uses of customer data must be confirmed in writing.
- Is there a SOC 2 report or ISO 27001 certificate?
- The website makes no claim to either. No report or certificate is supplied in this pack.
- Where are hosting and provider details?
- Customer hosting locations and an approved provider list are not published here. Your service documents should state who receives data, where and why it is processed, who can access it, how long it is kept and how it is deleted.
- Can we restrict AI providers or use an existing agreement?
- Customer-approved provider and endpoint restrictions are an architectural requirement. No customer model-provider route or bring-your-own-cloud deployment is confirmed here. Before an evaluation, verify the permitted models, contract, data region, retention, logging and failover behaviour for the exact service.
- Who owns generated material, and can we export everything?
- Ownership, licence rights, permitted export formats and data return must be agreed in writing. Ask which study versions, sources, issue records, decisions and audit records can be returned. Third-party source rights can limit redistribution; a displayed record is not automatically exportable.
- What must a human approve?
- Responsible specialists retain material clinical, scientific, statistical and regulatory decisions. The planned issue workflow distinguishes an AI candidate, a human-created issue, review, resolution and protocol approval. A completed readiness check cannot supply that approval.
- How would AI activity and decisions be audited?
- The planned record links the authorised context, source and method versions, proposed changes, validation, responsible human and rationale. Verify coverage, access, retention and export against the exact evaluated release. The public examples retain session actions only.
- How do roles and permissions protect our study?
- Access must be scoped to the organisation, study, role, object, field and permitted purpose. A contributor or AI model must not gain access through a wider backend credential. Private sponsor analysis and competing CRO material need separate verified protections; the public demo does not prove customer isolation.
- What is validated today versus planned?
- The public demonstrations show fictional, prepared examples. Protocol Integrity, formal issue handling and readiness enforcement are planned capabilities, not released customer controls. Check the capability register, exact product version and supporting security, privacy and clinical qualification before agreeing use. Review current capability scope →
- Who do we contact?
- Contact Tenvia for non-confidential enquiries. Support hours, response commitments, incident escalation and the legal name of the company providing the service need to be specified for your engagement.
Document availability · Reviewed 1 October 2026
See what is available
before requesting a pack.
Open the public resources below. The other entries show documents whose availability still needs confirmation. We only describe a document as available under a confidentiality agreement when we have confirmed that it exists and can be shared.
Available now
Supplier review checklist
A guide to planning your checks. It does not prove that the listed safeguards are in place.
Open the checklist →Available now
Evaluation scorecard
A blank worksheet for a fair comparison of quality, time and unresolved questions.
Open the scorecard →Available now
Example decision record
Fictional RD-204 review; not a customer export or validated audit record.
Inspect the record →Availability not confirmed
Hosting and third-party data and AI providers
An approved provider list for a customer service is not published. Ask where data goes, who can access it, how long it is kept and whether it is used for AI training.
Request availability →Availability not confirmed
Security architecture and independent testing
Public principles are available; the existence, date, scope and disclosure conditions of assessment reports need confirmation.
Request availability →Availability not confirmed
Quality, validation and release evidence
Ask what the product is intended for, its version, test results, known limits and the responsibilities of Tenvia and your team.
Request availability →Availability not confirmed
Confidentiality, service and data-processing agreements
The website describes the contracting path. Approved agreement versions, legal entities, pricing and service commitments are not published.
Request availability →Availability not confirmed
Recovery, support and incident arrangements
Support levels, recovery times and incident-reporting commitments are not published here. Obtain these terms before customer use.
Request availability →Not published
Customer outcomes and clinical performance evidence
This buyer pack does not include customer case studies or measured clinical performance results. Public demonstrations use fictional data.
Request availability →The review date refers to this list, not to an audit. For any report, check who issued it, what it covers, its date, relevant product version, limits and sharing conditions.
Your protocol and your data
What happens to your protocol?
These are the steps to agree before an evaluation. They are a planning guide, not a confirmed description of a live customer service.
- 01
Agree what enters
Define permitted documents, users and data handling before sharing a protocol.
- 02
Check the study structure
Review extracted information against its source. Confirm, correct or reject it.
- 03
Examine the evidence
Check permitted sources, relevance, dates and gaps alongside the design question.
- 04
Review AI assistance
Agree model providers, permitted inputs and data routes. Check suggestions and their limitations.
- 05
Make the human decision
Responsible specialists examine the proposal. The team records the decision and its reasoning.
- 06
Preserve or return the record
Agree version history, access, export, retention and deletion for the engagement.
Where would benchmarking data come from?
The planned workflow uses registered and permitted sources such as public registries, accessible protocols, amendments and literature where source-specific access is qualified. ClinicalTrials.gov and AACT copies of one NCT study must retain their source history and must not count twice. The website comparison uses 42 fictional studies, not a live dataset. No named-source operational access is promised here.
Can we inspect how a cohort was selected?
A released comparison must show study identities, filters, match criteria, method/version, denominator, source snapshot and retrieval dates, exclusions, missingness and applicability limits. Historical similarity does not establish clinical appropriateness, statistical adequacy or recruitment feasibility.
How are facts separated from AI suggestions?
Planned intake retains the original document/version and page or section beside the fact, extraction limits, AI proposal and attributable human correction. A confirmed draft is not an approved protocol version. Verify this behaviour for the exact evaluated release.
How are private sponsor documents isolated?
The required model separates public evidence from sponsor-private documents and derived results. Organisation, study, object, field, blinding and disclosure permissions must apply to retrieval, summaries, exports and caches. A CRO working for multiple sponsors has no pooling rights by default. Deployment evidence must prove these boundaries before confidential use.
What happens when information is missing or specialists disagree?
Missing evidence, unreadable material, stale reviews and unassessed dependencies must remain explicit. Do not infer a reassuring score or agreement from silence. The intended record preserves dissent, role-specific consequences, responses and the human disposition against the exact version.
How are versions and output changes traced?
Planned outputs pin the study/baseline, template, mapping/method and source/decision versions. A change identifies affected artifacts and prepares new drafts for review. Earlier approved artifacts remain intact; generation, export, sharing and release are separate permissions.
How would downstream transfers be validated?
Each future route must qualify the exact approved study, destination product/version/environment, mapping, permissions, compatibility tests and semantic read-back. Partial results or uncertain acknowledgements are unresolved, not successful. An assisted non-production Rave ALS draft-import route is planned; no live or universal EDC connector is promised.
Request a diagram showing where your data goes and a list of the AI providers for your service. The proposed system includes a gateway to control access to AI models. Ask for evidence that it works and protects data in the service you will use.
Explore AI and model governance →Review by topic
Go deeper by review area.
Security, privacy and data governance, AI, clinical quality and procurement each have their own questions and evidence list.
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 reviewConfidentiality agreement and approved access
Tell us which security, privacy, quality or architecture evidence your team needs to review.
Your review
02 / Describe your prioritiesPlease 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?