How we use AI, in plain language. 

We access only the systems and records needed to build and run the systems you have engaged us for — typically your call logs, CRM records, documentation, and the specific documents a workflow processes.

Access is scoped per engagement, listed in writing before work begins, and revoked at the end of the engagement or on request.

Client data is processed by the model and platform providers named in your engagement, under business or enterprise terms that exclude your data from provider model training wherever those terms are available.

We will tell you which providers are involved, and in which regions your data is processed, before you sign.

Every system we build has defined confidence thresholds and escalation rules. Low-confidence extractions, ambiguous requests, and anything the system was not designed to handle route to a person.

Content produced by our content engine is reviewed by a human editor before publication.

Deception or impersonation of a real person without disclosure. Undisclosed AI agents where disclosure is legally or ethically required. Scraping in breach of terms or law. Spam or outreach outside CAN-SPAM, CASL and GDPR requirements. Surveillance of employees or customers beyond lawful, disclosed business monitoring.

Voice and chat agents we build identify themselves as automated assistants when asked, and always offer a clean route to a human.

Recordings, transcripts and processed documents are retained for the period set in your engagement and deleted on request, subject to your own legal retention obligations.