AI transparency

You should know when AI is involved and what it is doing. This page sets out how Po uses it, on this site and in client work.

The models we use

• Our live demos run on Anthropic Claude. If it is unavailable, Google Gemini is used as a backup. Each demo result names the model that produced it.
• For client systems, we choose the model to fit the job and name it in your agreement.

Honest traces

Every demo shows a step-by-step trace marked ran, simulated, skipped or failed. We never present a simulated step as real, and we never report success when something failed.

Where a person stays in control

• Messages to customers, payments, deletions and other actions that are hard to undo wait for a person to approve them by default.
• Customer-facing answers come from your own approved knowledge, and questions it does not cover go to a person.
• Every automated run is logged so you can see what happened and when.

Limits

• AI output can be wrong, incomplete or out of date. Review it before relying on it.
• The demo models have no tools and take no actions outside this site.
• We do not use AI for legal, medical or financial decisions without a qualified person in the loop.

Your data

Demo text is not stored by Po. Please do not paste confidential or personal data into a demo. See the privacy notice for full details.