AI Vendor Directory · Customer support AI
Does Intercom Fin train on your data?
Yes — trains on your data by default. Vendor-stated default: yes — anonymized fine-tuning by Intercom is on by default (all Fin plans), with opt-out. Intercom's Fin trust page states: 'Fin can use anonymized customer data for model fine-tuning, and you can opt out at any time with data deleted within 30 days.' Third-party LLM providers (OpenAI, Anthropic) are a separate question: they 'are contractually restricted from using your data for training, with zero data retention policies in place once an output is generated.' So the vendor-side default to flag in reviews is Intercom's own anonymized fine-tuning — opt out via Intercom if policy requires. EU workspaces get Fin processing within Europe; AU workspaces' Fin processing is currently US-based.
By the Shieldra Compliance Team · Last updated 2026-07-29
What is Intercom Fin?
AI customer-support agent that answers end-user questions from your help content and conversation history, resolving tickets autonomously inside Intercom (also deployable on Zendesk/other helpdesks).
Key facts
- Vendor: Intercom — https://fin.ai
- Trains on customer data by default: yes — anonymized fine-tuning by Intercom is on by default (all Fin plans), with opt-out
- Source for the training answer: https://fin.ai/trust-reliability
- Data residency: US; EU; AU (workspace hosting; Fin AI processing currently US-based)
- Last verified: 2026-07-29
What certifications does Intercom Fin hold?
- SOC 2 Type II
- ISO 27001
- ISO 27701
- ISO 27018
- ISO 42001
- HIPAA
Security, DPA, and subprocessor links
- Security / trust page: https://trust.intercom.com/
- Data processing agreement (DPA): https://www.intercom.com/legal/data-processing-agreement
What is your EU AI Act role when you build on Intercom Fin?
Deployer under the EU AI Act, and Article 50 transparency applies: end users chatting with Fin must be clearly informed they are interacting with an AI system — keep Fin's AI identity disclosure enabled and route to humans on request.
What to record in your AI registry
- Provider: Intercom
- Model type: third party api
- Data typically flowing to the vendor: customer conversations, end-user PII, help-center/knowledge content, ticket metadata
Sources
- https://fin.ai/trust-reliability
- https://www.intercom.com/help/en/articles/7837535-fin-ai-agent-faqs
- https://www.intercom.com/legal/data-processing-agreement
Disclaimer
Curated from vendor-published pages on the date above. Vendor terms change - verify against the cited sources before relying on a profile for a procurement decision. Last verified 2026-07-29.
Frequently asked questions
Does Intercom Fin train AI models on your data?
Yes — trains on your data by default. Vendor-stated default: yes — anonymized fine-tuning by Intercom is on by default (all Fin plans), with opt-out. Intercom's Fin trust page states: 'Fin can use anonymized customer data for model fine-tuning, and you can opt out at any time with data deleted within 30 days.' Third-party LLM providers (OpenAI, Anthropic) are a separate question: they 'are contractually restricted from using your data for training, with zero data retention policies in place once an output is generated.' So the vendor-side default to flag in reviews is Intercom's own anonymized fine-tuning — opt out via Intercom if policy requires. EU workspaces get Fin processing within Europe; AU workspaces' Fin processing is currently US-based.
What certifications does Intercom Fin hold?
Per the vendor's published pages as of 2026-07-29: SOC 2 Type II; ISO 27001; ISO 27701; ISO 27018; ISO 42001; HIPAA.
Where does Intercom Fin store your data?
Vendor-listed data residency options as of 2026-07-29: US; EU; AU (workspace hosting; Fin AI processing currently US-based).
What is your EU AI Act role when you build on Intercom Fin?
Deployer under the EU AI Act, and Article 50 transparency applies: end users chatting with Fin must be clearly informed they are interacting with an AI system — keep Fin's AI identity disclosure enabled and route to humans on request.
