Trust and security
Answers for your security review, in one place
How Lumen handles your data, who can access it, and what the agent is allowed to do on its own.
Data
How your data is handled
Plain answers to the four questions every review starts with.
- Where it lives
- Your data stays in the US or EU region you choose at setup.
- How long we keep it
- 30 days by default. Set a shorter window per agent, or delete on request.
- How it is protected
- TLS 1.2 or higher in transit and AES-256 at rest.
- Model training
- Your data is never used to train models. Not ours, not anyone else’s.
Permissions
What the agent can and can't do
Set what each agent can read, draft and send. Anything risky waits for a person to approve it.
Referral agent · Health team
5 rules
- Read intake forms and referral lettersAllowed
- Draft referral summariesAllowed
- Route referrals to a specialty queueNeeds approval
- Message patients directlyNeeds approval
- Edit records in the EHRNever
Audit log
Access and audit
See what the agent read, what it wrote and who signed off. Export the full record for any review.
Today · Health team
Export CSV
- 9:42Referral agent read 42 letters
- 9:44Draft summary created
- 9:51Summary approved and sent
- 10:03Alert #4471 marked benign
- 10:15Policy reply edited, then sent
Access
Only the right people, and a record of what they did
Single sign-on
SAML and OIDC with your identity provider. People join and leave through the tools you already manage.
Role-based access
Decide who can run an agent, who can approve its work and who can change its rules.
Audit log export
Every read, draft and approval, exported as CSV or sent to your SIEM.
Compliance
The documents your reviewer will ask for
We share current reports under NDA, and only list certifications we actually hold. Questions go straight to security@lumen.example.
Data handling overview
On request
Subprocessor list
On request
Penetration test summary
On request
Data processing agreement
On request
FAQ
Questions reviewers ask
No. Customer data is never used to train models, and that commitment is written into your contract.
Only where you allow it. By default, every action that leaves Lumen waits for someone on your team to approve it.
Yes. We sign a BAA for health teams and a DPA for any team that handles personal data.
Write to security@lumen.example. We reply within one business day and keep you updated until it is closed.