Data and Information Security
You can also visit our Trust Center to learn about our security posture and request access to our security documentation.
Do you use customer data for AI training and improvement?
No. Customer and user data is never used by Outdoo or its sub-processors for AI training or prompt improvement.
What data does Outdoo collect, and how is it secured?
Outdoo collects limited personal information required to provide the service, including:
- Name
- Email address
- IP address (for security and error logging)
- Recordings uploaded to Outdoo
All data is encrypted in transit using TLS 1.2+ and encrypted at rest using AES-256+ standards.
Outdoo maintains strong security and privacy controls, and all sub-processors are required to uphold the same level of protection.
What AI providers does Outdoo use, and how is customer data protected?
Outdoo primarily uses custom fine-tuned LLM models hosted in AWS Bedrock. External AI providers are used only in limited cases related to product development and research & development.
Outdoo works exclusively with enterprise-grade, secure, and SOC 2-compliant AI providers. The company applies the minimum available data retention settings and maintains enterprise agreements to ensure customer data is not used for model training.
Where are Outdoo’s servers located?
All data is processed and stored in AWS data centers located in the United States:
- US-East (Virginia)
- US-West (California)
For EU customers, recording data is stored in Frankfurt and/or Ireland.
How are user permissions and roles managed?
Privileged roles and permissions are assigned manually by administrators to ensure controlled and secure access management.
Does Outdoo support Single Sign-On (SSO)?
Yes. Outdoo supports Single Sign-On through:
- Google OAuth
- Microsoft OAuth
- SAML 2.0
- OpenID Connect (OIDC)
How does Outdoo ensure evaluation reliability and accuracy?
Outdoo uses internal evaluation datasets and testing processes to continuously improve output quality and reliability while leveraging state-of-the-art AI models.
What safeguards are implemented to address bias, hallucinations, and misuse in AI?
Outdoo implements multiple safeguards across bias mitigation, hallucination prevention, and misuse protection.
Bias Mitigation
- AI models are rigorously tested for bias before deployment and undergo quarterly audits to meet defined fairness metrics.
- Biases stemming from data imbalance or human cognitive bias are actively identified and mitigated.
- Diverse project teams help anticipate and reduce potential bias risks.
Hallucination Prevention and Reliability
- AI systems are monitored quarterly and during major releases to ensure outputs remain within acceptable accuracy thresholds.
- AI-generated outputs may include contextual information such as explanations or confidence indicators to reduce over-reliance on automated responses.
Misuse Prevention
- Strict input and output guardrails validate and sanitize customer-provided data to protect against malicious inputs.
- Automated content filters continuously monitor generative outputs and flag or block content that violates standards, including sensitive or inappropriate information.
- Internal policies and technical controls prohibit unauthorized AI usage and ensure only approved, vetted AI systems are used.
When is customer data deleted?
Data may be deleted under the following circumstances:
- When a user deletes a recording from their library
- When a user deletes their account
- When an organization’s retention policy automatically removes recordings after a specified period
- When agreements with Outdoo end, customer data is deleted within 60 days unless otherwise agreed
Are recordings shared by default?
No. All recordings are private by default unless explicitly shared by the user with external users.
Depending on organizational settings and permissions, recordings may be accessible to managers, administrators, or internal teams.
If you have any questions, reach out to support@outdoo.ai