In August 2025, we announced our intention to fine-tune and use our own Al models for Fin. We have developed a number of custom models, the Fin CX Model Suite, which are used to power various parts of our AI Products. This helps our AI Products deliver faster, more accurate answers for our customers. Customer data remains private and secure, and is used solely to improve model performance for our AI Products.

The Fin CX Models are trained using de-identified customer data (as described below) from production Fin conversation logs. Conversations between our customers’ human agents and their end-users are not used for training. All training data is anonymized and de-identified, meaning we remove direct and indirect identifiers of the end-user involved in the Fin conversation, as well as any customer-specific references or identifiers.

We automatically exclude the use of customer data for model training for certain customer types, as follows.

We understand that some customers don't want their anonymized customer data used for Al training and fine-tuning. For this reason, we offer an opt-out which Customers can request at any time by following the process outlined below. If you want to opt out of AI model training and fine-tuning, you can request an opt-out through Settings > Workspace > Security > AI model training and fine-tuning.

For customers on a negotiated contract your Relationship Manager will take you through the opt-out process, as well as supporting any additional questions you may have. Alternatively, you can submit your opt-out request via our legal alias (legal@intercom.io).

We will process all customer opt-out requests within 30 days of the submission of your request. Once the opt-out is granted we will not use your customer data for the creation of any further training data.

Note: By opting out you may not benefit from certain Fin performance improvements delivered by our Al models.

We use Customer Data for product improvement, as specified in Section 3.2 of our Terms of Service. Customer data includes the Input and Output of our AI Products (see section 2 of our AI Products, Additional Product Terms). We use anonymised customer data to improve our Services by training and fine-tuning custom AI models for Fin.

We only train our custom models on conversations where Fin is involved. Conversations between your human support agents and end-users are not used as part of our training data. Training data will typically include the content of the conversation between Fin and the end-user, conversation attributes (such as date of conversation, answer length) as well as any knowledge content/articles retrieved by Fin when responding to the end-user query.

The meaning of “anonymized data” can differ depending on the relevant data protection laws that apply to your business. When we refer to anonymized customer data, we mean that we will remove direct and indirect identifiers of natural persons, so that individuals cannot be identified from the training data. We also remove any customer specific references such as company name or address. De-identification is completed prior to any model training. We will continue to treat the anonymized training data as being subject to GDPR, and subject to our obligations to you as Processor of your Customer personal data.

All of the current fin-cx models are based on open source (weight) models and the training and hosting takes place within our AWS infrastructure.

We have a number of controls and independent assurance processes in place to support the responsible development and use of AI. Fin is ISO 42001 certified, which means our AI management system is independently audited against an international standard for the responsible development and use of AI. This covers how we develop, test, and release our AI products. Our ISO:42001 certificate is available here.

Fin is also independently tested every quarter by the Artificial Intelligence Underwriting Company (“AIUC”). Its testing covers areas including data privacy, model training, data access, personal data leakage, cross-customer exposure, and intellectual property protection. You can find more information about our AIUC testing and review the test reports here.

The following customer types are automatically excluded from AI model training:

Customers who were previously granted an opt-out will continue to have their customer data excluded from any AI model training data. This includes customers who have a Business Associate Agreement (BAA), and customers who are on our EU or AU regional hosting plans and trial accounts.

Yes, you can opt back in through your workspace Settings > Workspace > Security > AI model training and fine-tuning, or by contacting legal@intercom.io or your Relationship Manager.

No, customers aren’t able to select which AI models Fin uses to provide the Services. We use a mixture of models to ensure that Fin delivers the best performance possible for our Customers.