- Deployment
- hosted
- Hq Country
- United States
- Hq City
- San Francisco
- Ownership
- private
- Founded
- 2020
- Training Default
- trains-deidentified
- Residency Options
- United States
- Gov Access Exposure
- us-cloud-act
- Eu Ai Act Role
- provider
- Certs
- SOC 2 Type II
- DPA available
- yes
- Retention Default
- Personal information is stored for as long as needed for the purposes in the privacy policy or as required by law. The trust centre states what is collected in unusually plain terms: calendar data including attendee emails, meeting content data including the recording and transcript, and user data.
- Retention Exceptions
- The two training sentences sit in the same clause. "We do not authorize third parties to use your personal information or User Content to train their artificial intelligence models" is followed immediately by "We may use and create de-identified data generated from your User Content to train, customize or improve our in-house artificial intelligence models." Unlike Granola there is no opt-out published for the in-house training. Fathom is certified to the EU-US Data Privacy Framework.
- Subprocessors
- Google Cloud Platform (US, primary cloud, handles recording and transcript), Anthropic (US, handles meeting content), OpenAI (US, handles meeting content), ElevenLabs (US, handles meeting content)
- Subprocessor Count
- 4
- Transfer Mechanism
- sccs+dpf
- Eu Procurement Ready
- fail
- Eu Procurement Reason
- Fathom's trust centre is a model of clarity about who touches what: four sub-processors, each tagged with whether it handles meeting content and whether it is forbidden from training on it. All four are American, and all four handle the recording or the transcript. The company is certified to the EU-US Data Privacy Framework, which is a real Chapter V mechanism and is credited here. What the buyer should read twice is the training clause, where a promise that third parties will not train is followed in the same breath by a statement that Fathom trains its own models on de-identified derivatives of the same recordings, with no published opt-out. There is no EU residency option.
- Training Claim Basis
- stated
- Training Confidence
- high
- Residency Claim Basis
- stated
- Residency Confidence
- high
- Retention Claim Basis
- stated
- Retention Confidence
- medium
- Subprocessors Claim Basis
- stated
- Subprocessors Confidence
- high
- Residency Class
- non-eu-only
- Portability / Exit Path
- proprietary-only
- Subprocessor Jurisdiction
- non-eu
- Underlying model providers
- OpenAI, Anthropic, ElevenLabs
