Researcher
The Researcher digs into your knowledge graph and the web to surface what your team already knows — and what it should. It searches atoms, scans documents, queries external sources, and synthesizes findings into clear, cited summaries that become atoms in your graph.
What the Researcher does
When you assign a task to the Researcher, it combines internal and external sources into a single, grounded answer. It doesn't just search — it synthesizes. You get a structured summary with cited sources, and the key findings are automatically saved as atoms so every future agent (and team member) benefits from the work.
When it activates
- Before starting a project — "what do we already know about this?"
- Competitive research — "how do our competitors handle onboarding?"
- Market analysis — "what's the current state of this space?"
- Pre-design — "what have users said about this pain point?"
- Due diligence — "what are the key risks in this approach?"
What it needs
- A clear research question or topic as the task description
- Optional: specific sources to include (documents, URLs, atom types to prioritize)
- Optional: a target KR or decision to anchor the research
What it produces
- A structured research report saved as an artifact on the task
- Key findings saved as LEARNING atoms in the Wisdom tree
- Source citations for every claim
- A summary comment with the top 3–5 actionable insights
Example: competitive research for a growth experiment
You're planning a referral program. Before building, you want to know how competitors handle it and what's known to work. You create a task: "Research referral program best practices and competitive landscape" and assign it to the Researcher.
The Researcher:
- Searches your knowledge graph for any prior atoms about referrals, growth, or user acquisition
- Searches the web for recent analysis on referral program mechanics
- Queries your connected Google Analytics data for existing referral traffic
- Synthesizes findings into a structured report
- Saves the top insights as LEARNING atoms — immediately visible to the Planner and Coder
When the Planner decomposes the referral epic an hour later, it already has this context.
Best for
- Competitive research before building
- Pre-build context gathering ("what do we know about X?")
- Answering strategic questions grounded in evidence
- User research synthesis from interviews or feedback
- Market analysis for a new initiative
Sources it searches
- Your knowledge atoms in the Wisdom tree
- Connected documents
- The web — public research, news, analysis
- Google Analytics, if connected
- Prior task summaries and artifacts