What is Research Mode?

Research extends Knowledge Agents beyond one-shot answers. Instead of returning a single response, it creates structured, explainable reports that synthesize hundreds of sources, cite evidence, and flag gaps—always within the bounds of existing role-based permissions.

How it works

Research builds trust by showing its work at every step.

• Start a conversation with any Knowledge Agent.
• Toggle on Research Mode when your topic requires deeper analysis.
• Guru drafts a research plan—outlining the steps it will take and the sources it will check.
• You review and refine the plan before it executes, ensuring the scope is right.
• Once confirmed, Guru runs multi-source searches and compiles a comprehensive report.
• The output includes citations, key insights, and recommended next steps—all while respecting the user’s permissions.

Behind the scenes, asynchronous job execution handles the heavy lifting—making it possible to run complex, multi-source investigations without slowing down your teams.

Why it matters

Most AI tools give you quick answers but no transparency. Research Mode flips that model: every investigation is planned, evidence-based, and permission-aware.

Employees save time by turning days of digging into minutes of results. Team leaders and SMEs gain confidence that outputs are grounded in verified content, not black-box AI.

What you can do with it

Research Mode is built for the kinds of questions that go beyond Q&A:

• Ops leaders can analyze churn trends, revenue gaps, or operational bottlenecks.
• IT teams can investigate security incidents, tool adoption, or recurring issues.
• HR & People Ops can synthesize policy compliance across regions and spot onboarding or training gaps.
• Sales teams can pull together competitive insights and pricing rules to support complex deals.
• CX leaders can identify recurring support themes, sentiment trends, and product feedback.
• Product & Engineering can analyze defect patterns, release impact, or adoption metrics across tools.

Governance and control

Every report is:

• Transparent — shows planned steps and the evidence used.
• Permission-aware — respects existing role-based access.
• Grounded — solely leverages internal company context, versus random web searches

This ensures enterprises can run investigations at scale without compromising compliance or trust.

Outcomes

• Save time — Turn days of manual work into actionable reports.
• Increase confidence — Every output is verified, cited, and explainable.
• Improve governance — Reports are fully auditable and reviewable by IT and SMEs.
• Close knowledge gaps — Flag missing or outdated content where employees need more.

Learn more about…

• Knowledge Agents
• AI Agent Center‍
• MCP Server‍
• Integrations

‍

Note: Knowledge Agent Chat & Research is free while in early access