"Is my data used to train models?" was answered on three pages with a true statement about cross-customer benchmarks, which is an answer to a different question. The actual answer is published now, on the security page, the FAQ and the pricing page: no model is trained, fine-tuned or improved on your data, every model call is inference, and the model providers' own commercial and API terms say the same about their side. The same page also now says that every deal, account and contact carries a change history showing the field, the value before, the value after, when, and whether a person or the AI made the change.
Security and your data →
EmpireOS is US dollars only, and there is no mobile app: it opens in a phone browser and the everyday screens work, but dragging a deal between stages does not work by touch. Both are now on the list of reasons this might be the wrong product for you, beside the ones that were already there. Two more answers were added as well: exactly what an import brings across, which is accounts, contacts and deals and not custom fields or activities, and how canceling works, which is a button in Settings with no minimum term behind it.
Why EmpireOS →
The first takes the objection that decides most of these deals seriously: reps do not resist an AI CRM because of the typing, they resist being measured by it, and the piece names that trade rather than talking around it. The second is about audit trails, and it includes the time our own tamper check reported a break, which turned out to be our own snapshot writer rather than anybody tampering. We fixed the writer, left the broken row in place as the record, and wrote down why.
Learn →