Responsible AI principles

People stay in control

Our agents hand conversations to your team, with full context, when a request needs judgement, empathy or authority the agent doesn't have. Where AI drafts expert work, such as a customs classification, a qualified person reviews and signs it off.

Every engagement has a named partner who is accountable for what we ship.

We're honest about what AI can't do

AI is not the right answer for every workflow. We say so when it isn't, we scope what an agent owns end to end versus only triages, and we don't promise that every request will be resolved without a person.

Measured before it's scaled

We agree success measures with you before go-live, test agents against real conversations, and roll out in phases so problems surface while volumes are small.

Your data, your rules

Client data is processed for that client's project under the written agreement for it, including retention, purpose limits and the subprocessors involved. Your DPO or counsel sets the requirements; we design to them.

Compliance is designed in, not claimed

For outbound calling and messaging we document consent, suppression lists and escalation paths with you, in line with rules such as TCPA and DNC in the US, the GDPR and ePrivacy rules in Europe, and India's requirements. We don't give legal advice and we don't display generic “certified” badges.

Transparent with the people on the other end

Callers and customers should be able to understand they are dealing with an automated agent and reach a person. We build disclosure and escalation into conversation design and follow your policy and local law on how it is worded.