Agents automate repeatable product management work by combining your Castspells.io data, curated tool access, and specialized prompts. Use the Agents workspace to delegate research, planning, and follow-up tasks while keeping full visibility into what runs and why.
Open Agents from the main sidebar or press ? on any screen to launch contextual help. The floating help button on the Agents page opens this guide directly.
Agents orchestrate long-running or repeatable work. Each task you queue captures a goal for the agent to execute using the tools it is allowed to call (creating entities, updating records, running analysis, or researching on the web). Results are saved back to your workspace and the task log so you can audit every step.
once, daily, weekly, or monthly repeats when the workflow needs to stay current.pending - Task is queued for the next run windowrunning - Agent is actively working the jobcompleted - Work finished successfully; review the output summaryfailed - Run ended with an error; open logs to inspect what happenedThe Coordinator Agent triages broad requests and routes work to the specialized agents best equipped to execute. Use it when you are unsure which workflow to call or when you want one run to orchestrate multiple artefacts (for example, researching a market, drafting supporting tasks, and planning the next experiment).
Great for
Focuses on defining and maintaining scopes. It can create scopes, refresh mission statements, update imagery, and surface scope-specific insights.
Try prompts such as
Handles competitive research, market sizing, and trend tracking. The agent can run web research, generate market sections, and keep your existing market documents current.
Use it to
Creates and maintains personas and linked people. It generates persona sections (goals, behaviours, jobs-to-be-done) and updates existing records with new research.
Helpful requests
Owns your task backlog. It creates, updates, and prioritizes tasks and user stories so delivery stays aligned to research.
Common workflows
Manages validations and insights. It can set up validation plans, log new findings, and synthesize insight summaries tied to your hypotheses and personas.
Ask it to
Keeps assumptions and hypotheses sharp. Use it to draft hypotheses with measurable signals, link evidence, and recommend next validation steps.
Prompt ideas
Supports experiments and trials. It drafts experiment briefs, updates running trials, and summarizes research outcomes so you can close the loop quickly.
Go-to tasks
search_entitiesQueue your first task with the agent that matches your current goal, then monitor the run from the Agents list. When you need guidance later, click the help button to jump back to this playbook.