Responsible AI.

ant. works alongside regulated teams in lending and payments. These principles guide how we design, build and deploy it.

1. People make the calls

ant. surfaces problems and proposes actions. Anything consequential, such as holding payouts, launching an offer or filing with a regulator, waits for a person to approve it.

2. Same access as the person

ant. can only see and do what the person it works for is permitted to. It never widens its own access.

3. Every action traceable

Each action is recorded in an audit trail: what was done, for whom, when and on whose approval.

4. Grounded in your sources

ant. shows the evidence behind its answers, such as the policy, ticket or record it relied on, so people can check its reasoning.

5. Your data stays yours

ant. runs inside your environment. We do not use one customer’s data to train models for another.

6. Fair and compliant

ant. does not make final credit decisions about individuals. We test for unfair outcomes and support customers’ obligations under fair lending, consumer protection and data protection rules.

7. Tested and monitored

We evaluate ant. before each release, monitor it in use and fix the issues we find.

8. Raise a concern

If you have a concern about how ant. behaves, email founders@theanthill.ai. We review every report.