AzureFeeds

Copilot helps every team make sense of their work — across chats, tickets, docs, and decisions scattered in tools like Jira, Confluence, and ServiceNow. It brings everything together through Bizchat, Search, and Declarative Agents, delivering precise, context-aware answers grounded in your organization’s data.

It knows where to look, understands what matters, and brings back insights you can act on with confidence. You can ground prompts in specific files or URLs, build scoped agents with clear knowledge boundaries, search naturally in plain English, and retrieve exactly what’s been shared with you.

Whether you’re debugging, researching, or writing, Copilot connects to your data in a way that feels more relevant, precise, and human — helping every user find what matters, faster.

1. Get Relevant, Source-Specific Answers with Content Source Filtering

Copilot draws from a wide range of sources — Outlook, Teams, Jira, Confluence, and more — to deliver comprehensive answers. But sometimes, users need more control to cut through noise and focus on verified information.

That’s why Copilot lets you decide where it looks. You can focus responses on trusted sources like Azure DevOps for open issues, Confluence for documentation, or Outlook for email threads — ensuring every answer is accurate, relevant, and grounded in the right context.

Example:

“Give me all the tickets assigned to me that are active/new.”

By applying the Jira filter, Copilot excludes noise from chats and emails and brings back verified data straight from Jira tickets.

 The result: more precise, authoritative answers that help you act faster.

Content Source filter in Researcher

The Copilot Researcher agent is an intelligent assistant built into Microsoft 365 Copilot, designed to handle complex, multi-step research workflows with precision. Through connectors, it securely accesses information from all data sources approved by the tenant administrator, empowering users to perform research across their organization’s complete knowledge ecosystem.

 For instance, researcher can create a process diagram and decision tree basis troubleshooting guides. Users can use it to create a decision-tree based step-by-step guide for setting up VPN to access company resources. By scoping the query to authoritative sources like ServiceNow users can ensure that Researcher is only picking up information from official sources.

Example  : “Create a decision tree style step-by-step guide to set up a VPN connection for accessing company resources.”
 (Source: ServiceNow Knowledge Base)

 2. Ground Prompts on Specific Files, Work Items, or URLs

Sometimes, the context you need lives in one specific item — not an entire data source. You can now ask Copilot to ground its responses on a specific third-party item, whether that’s a Google Drive file, a ServiceNow ticket, or a Jira issue.

There are two simple ways to do this:

1. Paste the URL from a connected data source.

Copilot is built to deliver responses grounded in trusted, authoritative content, and one of the most intuitive ways to define that context is through URLs. In everyday collaboration, users often share links to documents or items. Users can simply paste a link from a connected third-party data source, and Copilot will automatically resolve the URL to its indexed content, interprets the content behind the link , respect access permissions, and generate grounded, citation-rich answers.

Example: “Summarize this document [Google Drive link].”

2. Context IQ in Copilot chat ( / command) – CIQ allows users to quickly discover and reference relevant content from connected data sources using intuitive shortcuts—such as forward-slash commands or the Others tab. Copilot then grounds its responses in the actual indexed content, honoring access permissions and delivering citation-rich, trustworthy answers.

3. Find What Matters — Harnessing Third-Party Data in Copilot Search

Copilot Search helps users tap into this broader ecosystem by surfacing relevant results across both Microsoft and external data sources. Powering informed decisions and context-rich responses.

Right Rail in Copilot Search

The right rail of Copilot Search is built to enhance the search experience, helping users quickly find the data they need. Users can choose from various data sources—like Outlook, Teams, Jira, or Google Drive—so results come only from their selected source. This customization lets users refine their searches based on their specific needs. Additionally, the right rail displays which data sources are included in Copilot Search.

Filters in Copilot Search

Copilot Search offers a comprehensive set of filters that make it easy to find exactly what you need. Copilot Search has global filters which is You can refine results across all data sources by content type (documents, emails, tickets), date range, relevance, or person—helping you quickly locate work authored, assigned, or mentioned by specific colleagues.

When you narrow your search to a particular data source, Copilot automatically surfaces filters relevant to that platform. For example, in Jira you can filter by project, sprint, or ticket status, while in Teams you can filter by channel or message type. These global and source-specific filters work together to deliver results that are both relevant and deeply contextual to your workflow.

4. Create Focused, Relevant Agents with Scoped Knowledge

Copilot agents with third-party connector filters seamlessly integrate with platforms like Jira, Confluence, and ServiceNow to deliver organization-specific insights. With Scoped Knowledge, creators can now restrict each agent to defined segments of connected data — ensuring responses are focused, relevant, and trustworthy. Instead of searching across an entire platform, agents can be scoped to a single Jira project, Confluence space, or ServiceNow Knowledge Base for precise, context-aware results.

Example:

Arjun, a Project Manager, leverages an agent scoped specifically to the Jira project space for the “Mobile App Revamp”.

This agent goes beyond simple prompts, offering tailored responses for complex team queries.

For example, Arjun can ask:

• “Show all open critical bugs assigned to me.”


• “List all completed user stories this month, grouped by assignee.”


• “Display progress of sprint tasks and list all the bugs sprint wise.”

The agent references only verified issues and tickets within the scoped project, ensuring information is always relevant and up-to-date.

5.  Conversational Search with Context & Identity — Ask Naturally, Get Structured Results

We’re making structured data feel natural.

 With Conversational Search, you can query sources like Jira or Azure DevOps in plain English — and Copilot translates your request into the right filters automatically.

 Whether you’re looking for issues resolved last week, tasks assigned to a teammate, or work items tagged with a specific priority, just type your question the way you’d normally ask it. Our system intelligently interprets your query, identifies the relevant fields and filters, and brings back the exact results you need—no complex queries or manual filtering required.

This feature bridges the gap between natural language and structured data, allowing everyone on your team to find insights effortlessly, even without technical know-how. By reducing the time spent configuring filters, it helps teams stay focused on decision‑making and delivery. Try it out to experience a more intuitive, conversational way to explore your work data.

Let’s see this in action with an example. John is a Product Manager working with Contoso to onboard them to his product. His team uses Azure DevOps Work Items (ADO) to manage their work and progress. John has an important customer meeting with Contoso scheduled tonight. He wants to prepare by reviewing work items relevant to this customer, identifying any risks, and crafting an update.

John goes to Copilot and makes the following query

User query: “I’ve a customer meeting tonight. Can you fetch all work items tagged ‘Contoso’?”

Copilot chat can respond with up to 50 items per query and for this query, it returns 14 results, which includes work items of different priorities and states. This allows John to get an overview of all the items he should look at without having to deal with multiple queries and boards.

John then decides to narrow his focus to P0 or P1 items which are due soon as that will be the topic of conversation with the customer.

User Query: Of the above work items, how many are P0 or P1 and are due in the next week? Also, can you group them by state?”

John sees that there are 3 closed, 2 resolved and 1 Active item. The responses are grouped by state. John is now happy that most of the high priority issues are now resolved, and he will now have some good news to deliver to Contoso.

He now wants to understand more about the one active issue and he decides to drill down into it to understand the latest update so that we can provide an update to the customer.

User Query: “Ok. Can you summarize the latest updates on the Active work item above & help me prepare an update for my customer meeting tonight?”

Copilot can parse and understand the complete context of your work items across Jira, Azure DevOps, and other ticket-based systems—not just fields, but also comments, discussions, and attachments. This allows John to access the latest updates, decisions, and file contents tied to each ticket for a truly comprehensive view of his work and he sees the latest update right from Copilot without having to navigate into individual items.

6. Find What Others Shared With You — Now Identity-Aware

Collaboration just got smarter. Copilot can interpret sharing activity across connected systems and understand who shared what with you — even when user identities differ across tools.

Copilot chat can understand user-to-user sharing activity across tools like Jira and Azure DevOps. When teammates share documents, tickets, or other items with you, simply ask in natural language—
 “Get work items shared with me by Vivek” or “Show design docs Priya shared last week”—and Copilot will instantly surface the right items, without the need to search through emails, notifications, or tabs.

Behind the scenes, Copilot now performs cross-system identity resolution, mapping user identities from third-party tools to those in your organization. So when you say “John,” Copilot knows exactly which John you mean and retrieves accurate, complete results.

The result: a faster, more reliable way to find shared work, maintain momentum, and collaborate confidently—knowing Copilot always points to the right people and the right content.

Sample Query: Can you get the ADO items shared with me by Vivek?

The Bigger Picture: Clarity Through Connection

Together, these updates bring a simple but powerful shift — Copilot no longer just “searches everywhere.”

 It knows where to look, understands what matters, and delivers grounded, context-aware insights from across your organization.

• Ground prompts directly in files, items, or URLs


• Build scoped agents with precise knowledge boundaries


• Search naturally in plain English


• See the full picture with comments and attachments


• Retrieve exactly what others shared with you

Whether you’re debugging, researching, or writing a summary, Copilot connects to your data in a way that’s more relevant, precise, and human.