Achilles Documentation
Achilles helps banks, insurers and financial-services teams put AI agents in front of customers and staff — agents that answer questions, look up policies and accounts, and take actions, within rules your risk team controls.
You don't need to write code. You describe the job in plain language, connect the systems the agent may read, and Achilles builds, tests and runs the agent for you. Every conversation is recorded and every action is checked against your rules.
Start here
| I want to… | Read |
|---|---|
| Build my first agent in about an hour | Quickstart |
| Understand the words we use | Key terms |
| See how Achilles works, end to end | How it works |
| Turn a process document into an agent | Agent Builder |
| Know how customers' data is protected | Safety overview |
| Put an agent on our website | Website widget |
| Hand this to our IT team | For IT teams |
What you can do with Achilles
- Answer questions about policies, claims, loans, renewals and products — using your own data, not guesses.
- Take actions such as raising a service request or updating a contact detail, with human approval where you want it.
- Hand over to a person when a customer asks, or when the situation needs one.
- Stay in control with role-based access, personal-data masking and a full audit trail.
Find your way around
- Get started — your first agent, key terms.
- Build agents — the Agent Builder, your process document, data, abilities, roles, voice, testing.
- Safety and control — guardrails, approvals, personal data, human handoff, audit.
- Launch and run — website widget, messaging channels, users and teams, monitoring.
- Console guide — a tour of the console by task.
- For IT teams — integration, API, models, self-hosting.
- Help — FAQ and troubleshooting.