How Rillet Releases Changes Faster

Rillet makes software that helps companies do their accounting work. Customers sometimes ask the company to change something in the software. The details used to be spread across several tools and conversations. An engineer needed time to put those details together before starting. Rillet changed this process so more information could travel with each request.

A customer support worker can now highlight something on the screen. The worker then writes what the customer would like to change. An AI agent collects both pieces of information about the request. An agent is a tool that can perform tasks using AI. If the request is clear and safe, another agent prepares the change.

The agent can test the change and attach pictures of the result. An engineer looks at the work before it can go out. If the engineer asks for changes, the agent can continue the work. This keeps people in charge of what reaches customers. It also lets them spend less time finding the original request.

Rillet uses the same basic setup for agents across different teams. People can write instructions for the agents without learning to code. The company says its rate of releasing work tripled in three months. Some changes requested by customers have reached them within two hours. The example shows how clear requests and human checks can support faster releases.