GAN Integrity is bringing AI to your compliance and third-party risk management programs. Take advantage of these new features when you're ready.
GAN Integrity AI Capabilities
Features and capabilities available in the platform today
← scroll to see all columns →
| AI Feature | PRODUCT(S) | Benefits | Customer data | Data Storage | AI technology | Opt-In (*) |
|---|---|---|---|---|---|---|
| TPRM, COI, Donations & Sponsorships, Gifts & Entertainment, Investigations / Incident Management, Policy Management | Create individual visuals or entire dashboards from your compliance data, turn existing dashboards into narrative AI stories, and ask questions of your data in plain language. | Records and documents from the connected GAN Integrity modules that a user chooses to query or report on. | Processed through Amazon Quick in your primary data-centre region (EU or US). Your data is not used to train the models. | Amazon Quick (AWS) | Yes |
* To opt-in and take advantage of our new AI capabilities please contact your Customer Success Manager.
Future AI enhancements
Features and capabilities coming soon
← scroll to see all columns →
| AI Feature | PRODUCT(S) | Benefits | Customer data | Data Storage | AI technology | Opt-In |
|---|---|---|---|---|---|---|
|
AI Summarization
|
Third-Party Risk Management (TPRM) | Generates a structured summary of a third party and shows key risks read directly from the third-party record. Reviewers see a vendor’s risk profile at a glance without having to read every underlying document. | Third-party records, questionnaires and screening results selected for summarization. | Content is processed on demand through AWS Bedrock in your primary data-centre region (EU or US). Inputs are not used to train the foundation models and are not retained by the model provider once the summary is returned. | AWS Bedrock, using managed foundation models. | Yes |
|
Form Prefill
|
TPRM vendor intake and disclosure workflows (COI, Gifts & Entertainment, Donations & Sponsorships) | Speeds up onboarding vendors. A user uploads a document or gives a plain-language description, and the form is prefilled with suggestions for the user to accept or reject. This reduces manual data entry, and can improve accuracy and completeness when working from a clear source. | Documents or plain-language descriptions the user provides to populate a disclosure or intake form. | Content is processed on demand through AWS Bedrock in your primary data-centre region (EU or US). Inputs are not used to train the foundation models and are not retained by the model provider once the form is prefilled. | AWS Bedrock, using managed foundation models. | Yes |
Frequently Asked Questions
Which sub-processors are used to provide the AI features?
Our AI features run exclusively within Amazon Web Services (AWS), including model inference. Please refer to our Sub-Processor list for more details.
What customer data is in scope for the AI features, and does this include personal data?
Three categories for services and support:
- The customer's workflow configuration blueprint is used to triage issues raised in support and bug tickets. This may include named individuals and email addresses.
- The content of support tickets themselves. Support tickets may contain personal data, such as email addresses of individuals named in or corresponding through the ticket.
- Relevant logs produced during the normal operations of the product in order to troubleshoot specific issues. This may contain identifiers for individuals in the customer organisation.
For features in the product which you may later explicitly choose to adopt, we will make it clear which data is in scope case by case.
Where is this data hosted and processed?
For product features, data is hosted and processed in the same location as the rest of the platform: EU or US as agreed for each customer. For service and support, the inference workloads are processed in the EU, where most of our personnel are also located.
Please refer to our Sub-Processor list.
Is customer data used to train or fine-tune AI models?
No. Customer data is not used to train or fine-tune any AI models unless otherwise explicitly agreed.
If data is transferred outside its country of origin, what is the legal basis?
Because the model is self-hosted within AWS EU regions and Anthropic does not process the data directly, there is no cross-border transfer to a third-party AI provider.
How long is data processed by the AI features retained?
Customer data is not retained after processing.
How is data protected while being processed by the AI features?
Data processed via AI features or AI-based support is subject to our standard policies surrounding data privacy. More information can be found on our Trust Center.