AI trust and security
AI trust and security
Licensing data is commercial data: rates, guarantees, partner performance. Here is exactly how it is handled when you use AI with Flowhaven, today and as our AI features grow.
Layer by layer
How your data is protected
Licensing data is commercial data: rates, guarantees and partner performance. Four layers keep it safe when you use AI with Flowhaven.
Today and next
From integration today to Agentforce next
| Today, through an integration | Next, with Agentforce on Salesforce | |
|---|---|---|
| How AI connects | Tools such as Claude connect to your Flowhaven data through an integration | Flowhaven AI features run inside Salesforce |
| Training on your data | Anthropic does not train its models on data from its commercial products | Prompts and responses are not used to train third party models |
| Data retention | Set by the AI provider’s business terms, which we help you check | Zero data retention with AI providers under Salesforce agreements |
| Sensitive data | The tool sees only what the connected user can see | Sensitive data can be masked before it reaches a model |
| Oversight | A person confirms every change | A person confirms every change, and every interaction is logged in an audit trail |
Salesforce Trust Layer
Built on the Salesforce Trust Layer
As Flowhaven AI features arrive on Agentforce, they inherit the Salesforce Trust Layer (formerly the Einstein Trust Layer). No other licensing platform is built on it.
Zero data retention
Prompts and responses are not stored or used to train third party models.
Data masking
Sensitive data is replaced with placeholders before a prompt reaches the model.
Dynamic grounding
Answers are grounded in your trusted, approved Salesforce data.
Toxicity detection
Responses are checked for harmful content before they are shown.
Audit trail
Every AI interaction is logged so it can be reviewed later.
Human in the loop
People stay in charge of decisions and of what goes out.
Our rules
How we use AI in licensing
Scheduled updates
Internal updates, such as a daily recap, can run automatically.
People send to partners
Anything addressed to a licensee is drafted by AI and sent by a person.
People confirm changes
AI proposes record changes. A person confirms them.
Legal reviews legal
Anything with legal weight goes to your legal team.
FAQ
We get these questions a lot.
Still have a question? Talk to our team, or see Flowhaven on your own program in a 30 minute demo.
Answers come from your own Flowhaven records, so every number can be traced back to an agreement, approval or royalty line. AI can still be wrong, which is why people approve before anything changes or goes out.
Answer quality follows record quality. AI flags gaps, such as unreconciled royalty lines or data that looks too perfect to trust, instead of hiding them.
No. Outputs with legal weight, such as customs letters, are evidence for your legal team to review, not legal advice.
Yes. AI is only used if you choose to connect an AI tool to Flowhaven.
See how Flowhaven connects to your systems
Enterprise integration, partner self-service, and open APIs. See how it fits your tech stack.
