Security
You are handing us every ticket. Here is how we treat it.
AI Autopilot reads ticket content and documentation across your entire client base. That deserves a straight answer about what happens to it.
73 controls · monitored by Vanta
- 20Infrastructure security
- 35Internal security procedures
- 5Product security
- 10Organizational security
- 3Data and privacy
AI processing
What the agents can and cannot do with your data
These are the commitments that matter most when an AI reads every ticket you handle.
Your data never trains a model
Ticket content, documentation, and client data are never used to train or fine-tune any model — ours or a provider's.
Stateless processing
AI providers do not retain your data after a request completes. Nothing persists on their side between tickets.
Credentials are never sent
Passwords, API keys, and secrets are blocked or masked before anything leaves your environment. Agents never see them.
Minimum necessary data
Only the data required to answer the request is sent. Agents do not receive a copy of your PSA.
Your permissions still apply
Agents inherit the access rules already configured in your PSA. They cannot see what a technician could not.
Graduated autonomy, with a human in the loop
Claudia starts supervised on every MSP. Your team sets the level she may act at and raises it over time, on evidence, with the tools we provide.
United States only
All data is stored and processed within the US, on Microsoft Azure. Nothing is transferred outside it.
Graduated autonomy
Every agent starts supervised. Your MSP decides how far it goes.
The current launch is supervised. The broader operating model describes graduated autonomy; any later expansion requires an agreed scope and evidence from the actual workflow.
- L0Claudia today
Advisory
Recommends only; a human does everything.
To reach L1Advisory recommendations are accepted (acted on) ≥80% of the time over a meaningful sample — the suggestions are trustworthy enough to draft from.
- L1Claudia today
Copilot
Drafts/acts on request; a human approves each step.
To reach L2~200+ copilot actions approved with ≤5% correction rate over ~4 weeks — the human is rubber-stamping, not rewriting.
- L2Claudia today
Supervised
Acts within scope; a human supervises + can intervene.
To reach L3≥6 weeks supervised with no intervention-worthy error and the quality metric at/above threshold — scope is well-bounded and the failure modes are understood.
- L3Permitted, on evidence
Guardrailed
Acts autonomously inside hard guardrails + budgets.
To reach L4Sustained in-guardrail operation under outcome (not step) review, with no guardrail breach and stable quality — humans review results, not actions.
- L4Not permitted
Autonomous
Runs end-to-end; humans review outcomes, not steps.
0 of 20
Agents cleared to act without a human
L2
Highest level any agent runs at today
The levels and the rungs are the operating model's, verbatim; the rung figures illustrate the shape of the bar rather than measured telemetry. Every agent starts supervised on every MSP, your team raises the level on evidence, and the highest level any class of agent is permitted today is L3. No agent is cleared to act past the checkpoint your MSP sets — the two figures above are derived from the operating model, not asserted here.
Controls
Continuously monitored, publicly published
Monitored by Vanta and published live. The Trust Center reflects their current state, not a snapshot from an audit window.
20
Infrastructure security
- Production database access restricted
- Production network access restricted
- Remote access MFA enforced
- Encryption key access restricted
35
Internal security procedures
- Continuity and Disaster Recovery plans established
- Continuity and Disaster Recovery plans tested
- Configuration management system established
5
Product security
- Data transmission encrypted
- Vulnerability and system monitoring procedures established
- Control self-assessments conducted
10
Organizational security
- Code of Conduct acknowledged by employees and enforced
- MDM system utilized
- Visitor procedures enforced
3
Data and privacy
- Customer data deleted upon leaving
- Data classification policy established
- Data retention procedures established
73 controls in total, of which the above are a sample. Data is encrypted at rest (AES-256) and in transit (TLS 1.2+).
Our SOC 2 Type II audit is underway, with the report expected Q1 2027. The full, live list is at trust.aiautopilot.ai.
The guide · 10 pages · PDF
Graduated autonomy: the MSP owner's guide to L0 to L4
How a desk raises an agent from advisory to supervised on evidence, and never past a human by default.
Retention
How long we keep things
On contract termination, customer data is deleted within 30 days unless the law requires otherwise.
| Data type | Retention period |
|---|---|
| Account information | Duration of account + 30 days |
| Billing and transaction records | 7 years (legal requirement) |
| Activity and audit logs | 90 days |
| Support communications | 3 years after resolution |
Full detail in the Privacy Policy.
Questions
The ones we get asked
Which AI providers do you use?
Anthropic Claude and OpenAI, hosted on Microsoft Azure within the United States. Both are configured so your data is not retained after processing and is never used for training.
Can we opt out of AI features entirely?
Yes. Contact us and we will disable AI processing on your account. The platform's non-AI functionality continues to work.
What happens to our data if we leave?
Customer data is deleted within 30 days of contract termination, unless a longer period is required by law or you request earlier deletion.
Do agents have access to our client credentials?
No. Credentials, API keys, and secrets are blocked or masked before any request leaves your environment. Agents operate on ticket content and documentation, never on secrets.
Are you SOC 2 certified?
Not yet — our SOC 2 Type II audit is underway, with the report expected Q1 2027. In the meantime our 73 security controls are continuously monitored by Vanta and published live in our Trust Center, covering infrastructure, internal procedures, product, organisational, and data privacy controls. Ask us for the report once it is issued.
Security review or questionnaire to complete? info@aiautopilot.ai
What the agents actually read is set out stage by stage, each one has its own page, and the connectors we hold credentials for are listed in full.
Bring your security questionnaire to the demo
We would rather answer it up front than at contract stage.