AI agents

Give an agent a role, connect its projects and build an executable action map with a clear start condition.

Turn live project evidence into a structured health review, practical recommendations and owner-controlled changes.

Receive useful project attention without turning the agent into a noisy duplicate project manager.

Understand which settings shape the agent’s responses, which define execution and which govern access.

Select a connected account and resource, then map its output into the next step instead of re-entering identifiers.

Give one saved agent a combined view of all or selected projects and processes, with its own action map, schedule and private reports.

Turn a responsibility into an executable workflow, answer concrete setup questions and connect real sources, actions and approval paths.

Add any working rule, preference or useful context in chat or Settings; review related guidance and control what the agent uses in future work.

Configure a report source, recipient action and schedule, then confirm the run and delivery instead of relying on the chat response.

Choose personal or company scope, inspect usage by tool, agent and provider, and distinguish Collty usage value from provider cost.