Prepare your team conversations with the Researcher agent

✅ How to

1. 1Why Researcher? Researcher is a deep-reasoning agent in Copilot M365 Premium. Unlike a regular chat reply, it runs a multi-step investigation and taps into Work IQ, your emails, meetings, Teams chats and documents, to produce a structured, nuanced and sourced analysis. Best when context and quality matter more than speed: exactly the case when preparing an important human conversation.
2. 2Open Copilot M365, then pick the Researcher agent from the agents list. Ask in natural language. Key point: before starting, Researcher asks you scoping questions (scope to cover, main objective, level of formalisation, sensitive items to include or not). Answer them to sharpen the result, or simply say “go ahead” to let it build with its best judgement.
   
3. 3Annual review: ask Researcher to gather the person's achievements, contributions and growth areas over the year, drawn from your exchanges and tracked projects, with dated, sourced highlights.
4. 41:1: have it prepare an agenda grounded in reality: follow-up on last meeting's actions, ongoing topics, warning signs (workload, blockers) spotted in recent exchanges, and open questions to ask.
5. 5Constructive feedback: go beyond a simple rephraser. Ask Researcher to analyse a specific situation from several angles (context, observed facts, impact) and propose balanced, actionable feedback using the SBI method, with concrete examples from the exchanges.
6. 6Team syncs & objectives: ask for a synthesis of the week's progress and blockers per person/project, then rewrite objectives into measurable OKRs and flag dependencies or risks.

Act as my Researcher agent. Using Work IQ (my emails, Teams meetings and documents from the last 12 months), prepare my annual review with [person]. Structured deliverable: 1) major achievements and contributions (with dated facts and sources); 2) evidenced growth areas; 3) balanced SBI-method feedback on [specific situation]; 4) 5 open questions for the conversation; 5) proposed objectives rewritten as measurable OKRs. Separate facts from assumptions and cite your sources.

💬 Comments

No comments yet. Be the first!

✨ Discover more

NewPremium

The 1P agents of Microsoft 365 Copilot: the full tour

You know the generalist Copilot. But did you know Microsoft already ships a small team of ready-to-use specialists? These are the 1P (first-party) agents, « Built by Microsoft »: Analyst, Researcher, Plan My Day, Workflows… Here's who does what, sorted by family, with concrete examples.

END USERIT PRO7 min read

NewPremium

Build a Copilot agent for a business use case

With Agent Builder, anyone can create their own agent, no code: describe it in natural language, add your instructions, enable options (images, sources) or start from a ready-made template.

IT PROEND USER5 min read

NewPremium

Create an Outlook rule with Copilot

No more digging through Outlook's rules menus: tell Copilot what you want to automate in plain language, and it creates the rule for you. Here's the principle, several concrete video examples, where to find your rules, and the few limits to know.

END USER5 min read