What this page answers
- What data is sent to AI providers for classification and analysis workflows
- Which provider and key-ownership modes are supported
- How retention and training controls differ by provider and configuration
Current state (as of March 5, 2026)
Tero supports three AI deployment modes: hosted default provider path, bring-your-own provider credentials, and customer-controlled self-hosted provider routing.AI workflow data scope
AI workflows use data needed for classification and analysis tasks. By default, this includes control-plane-relevant context and telemetry samples required for the task, not unrestricted ingestion of all customer telemetry.| Topic | Baseline approach |
|---|---|
| Prompt scope | Task-scoped inputs for classification and analysis workflows |
| Data minimization | Inputs are bounded to workflow requirements |
| Model output handling | Outputs are used for product classification and recommendation workflows |
| Governance | Provider path and key ownership can be customer-controlled by deployment model |
Provider and key ownership modes
| Mode | Description | Typical use case |
|---|---|---|
| Tero-managed provider path | Tero-hosted default provider configuration | Fastest onboarding |
| Bring-your-own provider credentials | Customer API credentials used for provider calls | Customer-controlled commercial relationship with OpenAI/Anthropic |
| Self-hosted provider path | Runtime and provider routing controlled in customer environment | Maximum infrastructure and network control |
Provider data controls (external policy alignment)
| Provider path | Training usage baseline | Retention baseline | Zero-retention option |
|---|---|---|---|
| OpenAI API | API data is not used for model training by default (unless customer opts in) | Abuse-monitoring and application-state retention depends on endpoint/configuration (default policies apply) | Available for approved organizations on eligible configurations/endpoints |
| Anthropic API | Commercial/API data is not used for model training by default | Standard API retention baseline applies (commercial policy default) | Available by agreement for eligible enterprise API use cases |
| AWS Bedrock | Prompts/outputs are not used to train base models and are not shared with model providers | Bedrock service data-protection model applies | Private connectivity and customer-managed encryption controls available |
Deployment boundary
| Control area | Tero-hosted | Self-hosted |
|---|---|---|
| Provider account boundary | Tero-managed or customer-provided credentials | Customer-controlled |
| Runtime/network boundary | Tero-hosted boundary | Customer infrastructure boundary |
| Provider policy configuration | Configured per provider path | Customer-configured |
| Compliance posture | Tero controls + selected provider controls | Customer environment + selected provider controls |
Model and provider coverage
- Recommended frontier providers for production quality are Anthropic and OpenAI.
- Bedrock and additional provider paths are supported as deployment-dependent integration options.
- Additional open-source/local model paths (for example, Ollama-compatible) are possible, with quality validated case-by-case.
Recommended enterprise hardening profile
| Control | Recommended setting |
|---|---|
| Provider account boundary | Use customer-provided credentials or self-hosted provider routing for strict boundary control |
| Data retention mode | Use provider zero-retention mode where eligible and contractually enabled |
| Training controls | Keep provider training opt-in disabled for API data |
| Prompt scope control | Limit prompts to minimum task-relevant context and avoid unrestricted raw payload submission |
| Secret management | Store provider credentials in managed secret stores with scoped runtime access |
| Key lifecycle | Rotate provider credentials on a defined cadence and on any incident trigger |
| Network controls | Restrict outbound destinations to approved AI provider endpoints |
| Auditability | Log AI request metadata and outcome events without exposing sensitive prompt text in operational logs |
External references
- OpenAI data controls
- Anthropic retention and zero data retention FAQs
- Anthropic zero data retention scope
- Amazon Bedrock data protection
- Amazon Bedrock security/privacy overview
Evidence you can request
| Topic | Primary evidence |
|---|---|
| High-level data scope | Overview, Data Handling |
| Ownership split | Shared Responsibility |
| Identity and credential controls | Identity and Access |
| Encryption and key controls | Encryption and Key Management |