AI transparency

We use AI to help clients request changes in plain language and to speed up routine delivery where it is safe and appropriate. AI supports our work; it does not replace human judgment for structural, risky, or brand-sensitive decisions.

Copilot is embedded on your site so you can describe updates in your own words. Requests may be routed through preview, publish, or human review depending on risk and your agreement.

Not every request is fully automatic. We will tell you when something needs design, development, or a scoped project.

Content you send through the assistant may be processed by our AI providers to generate responses and, where applicable, proposed changes. We configure providers so client data is processed to fulfill your requests and is not used to train public models.

Retention and subprocessors are documented in your statement of work or maintenance addendum and in our privacy policy.

We use staging, approvals, and rollback paths where our pipeline supports them. Destructive or high-impact publishes may require explicit sign-off. Exact rules follow your contract.

AI outputs can be wrong or incomplete. You remain responsible for reviewing published content, especially legal, medical, or regulated claims. We are not liable for reliance on automated suggestions without review.