AI Development and Customer Data Policy

Purpose

This policy describes how AI-enabled functionality is designed, developed, and used within FlowFuse, both in the product and internally by FlowFuse team members.

Its purpose is to protect customer data, maintain trust, and ensure responsible, transparent, and secure use of AI technologies.

Scope

This policy applies to:

• AI functionality shipped as part of the FlowFuse product
• Internal use of AI tools by FlowFuse employees and contractors
• Integrations with third-party AI services

Guiding Principles

1. AI is assistive, not autonomous.
2. Humans remain accountable for decisions and outcomes.
3. Customer trust and data protection take precedence over experimentation speed.
4. AI usage must be intentional, scoped, and reversible.
5. FlowFuse does not train AI models on customer data or company data.

Internal Use of AI by FlowFuse Team Members

FlowFuse supports and encourages responsible use of AI tools by team members to improve development velocity, quality, and operational effectiveness.

When using AI internally:

1. Do not share customer data with AI tools except as explicitly permitted by this policy.
2. Use internal, synthetic, or publicly available data for experimentation whenever possible.
3. Review AI-generated output before relying on it or including it in the product.
4. Do not use AI tools to bypass security controls, access restrictions, or approval processes.
5. Credentials may be given to an AI tool or agent where the work genuinely needs it, provided all of the following hold:
   • Discuss it first. Talk the use case through with at least one other team member before you create or share the credential. If you cannot explain to a colleague what the agent will do and why it needs this access, that is your answer.
   • Document the discussion. Record who you spoke to, what the agent will be doing, the scope granted, and when the credential will be revoked. Open an Access / Permission Request issue in the admin repo on GitHub, the same place other access decisions are recorded. A verbal or DM conversation nobody can find later does not count.
   • Scope it to the job. Grant the narrowest access that makes the task possible. For a FlowFuse Personal Access Token, that means scoping it to the single team you are working in, not your whole account.
   • Read only by default. Only grant write access where the task genuinely cannot be done without it.
   • Rotate and revoke. Treat any credential an agent has held as spent. Revoke it as soon as the work it was created for is finished, and do not carry it over to the next task. Where a credential must persist, rotate it on the schedule set out in the Access Control Policy, at minimum annually, and sooner if the agent, tool, or vendor changes. Revoke immediately if you suspect it has been logged, cached, or otherwise exposed.
   • Never admin-scoped. Admin-scoped FlowFuse Cloud and Dedicated Instance PATs are the highest-risk case: an agent holding one can take destructive, account-wide actions. That is never the minimum an individual task needs.
   • Passwords, and any other secret that cannot be scoped or revoked, are still not to be shared with AI tools.

Apply the same data protection, security, and review standards to internal AI usage as to AI functionality shipped within the product.

Approved AI Tools

FlowFuse provides the following AI tools to team members:

• Claude (via Claude Team) — general-purpose AI assistant for writing, analysis, coding support, and research.
• Gemini (via Google Workspace) — AI features built into Gmail, Docs, Meet, and other Workspace apps.

AI features built into other approved tools (for example, summarization, drafting assistance, or transcription) are covered by the existing approval of the underlying tool and the principles in this policy.

Other AI tools may be approved on a case-by-case basis through the vendor approval process. The full list of approved vendors and software is maintained in Vanta and via approved Vendor & Software Checklist issues in the admin repo on GitHub.

Requesting a New AI Tool

To request approval for a new AI tool, open a Vendor & Software Checklist issue in the admin repo on GitHub. The request is reviewed for security, data handling, and budget before approval.

To request access to an already-approved AI tool, open an Access / Permission Request issue in the admin repo on GitHub.

Customer Data Usage

Customer data may be used with AI systems only under the following conditions:

1. Use customer data solely to provide the requested product functionality.
2. Ensure all AI processing of customer data follows existing access controls, logging, and security policies.

Customer data is not used for the following purposes:

1. Do not use customer data to train shared, public, or cross-customer AI models.
2. FlowFuse does not train AI models on customer data.
3. Do not use customer data for internal experimentation unrelated to a customer’s use case.
4. Do not use customer data to improve general-purpose AI model behavior.

Customer data remains owned and controlled by the customer at all times.

Internal Data vs Customer Data

1. Internal, synthetic, or anonymized data may be used for:
   • Prototyping and experimentation
   • Prompt development
   • Evaluation and testing of AI features
2. Do not repurpose customer data for internal AI development or testing, even if anonymized, without explicit approval.

Third-Party AI Services

Third-party AI providers may be used within the FlowFuse product only when:

1. Appropriate contractual data protection terms are in place.
2. The provider does not retain or reuse customer data for model training.
3. Data shared is limited to the minimum required to provide the feature.

Review new AI provider integrations prior to use and assess them for:

1. Data handling and retention practices.
2. Security posture.
3. Compliance and risk implications.

FlowFuse maintains a documented inventory of approved third-party AI services used within the product that may process customer data. This inventory includes references to each provider’s relevant data handling and security policies. The inventory is reviewed as part of the vendor risk management process and is available upon request.

Product Transparency

1. Clearly identify AI-assisted functionality in the product and/or documentation.
2. Do not present AI outputs as authoritative or decision-final.
3. Inform users when AI is involved and ensure they understand that human judgment is required.

Prohibited AI Use Cases

Do not use AI for:

1. Fully autonomous decision-making affecting customers.
2. Safety-critical or high-risk operational decisions.
3. Surveillance, behavioral profiling, or user scoring.
4. Legal, medical, or employment decision-making.

Review and Oversight

1. Require new AI-enabled features to undergo review that considers:
   • Data inputs and outputs
   • Customer impact and potential failure modes
   • Third-party dependencies
2. Restrict, modify, or disable AI functionality if risks or assumptions change.

Exceptions and Enforcement

1. Require explicit review and approval for any exception to this policy.
2. Violations may result in feature rollback or removal of integrations.