I want an agent to help with multi-step research, but I still need to control the scope and end point.

Enable an agent to assist with multi-step research while keeping the user in control of scope, direction, and completion criteria. The feature helps users move faster through complex research without losing oversight of what gets explored and when the work is considered done.

This feature provides an agent that can carry out multi-step research tasks while the user explicitly controls the scope and the end point. Users can define what questions to answer, what sources or boundaries to follow, and what a “finished” result looks like before the agent proceeds. The agent can then break the work into steps, gather and synthesize information, and report progress in a structured way. Throughout the process, the user can refine the scope, stop exploration of certain branches, or adjust the intended output. This helps prevent research from expanding indefinitely or drifting away from the original objective. The feature supports iterative research workflows where partial findings guide next steps, without requiring the user to manually manage every subtask. It is useful for comparing options, building background knowledge on a topic, preparing briefs, or validating claims across multiple sources. By keeping the end point user-defined, it improves predictability and reduces time spent on unnecessary investigation. Overall, it accelerates multi-step research while maintaining clear user oversight and decision-making.