On July 8, 2026, GlobalData Plc announced that its AI research analyst, Ava, is being integrated directly into Microsoft 365 Copilot. Ava will appear as a Copilot extension, letting business users query GlobalData's vast repository of market research, company profiles, and analytics without leaving Word, Excel, PowerPoint, or Teams. The integration uses the Model Context Protocol (MCP), an open standard that connects AI models to external data sources, to bring proprietary intelligence into the flow of everyday work.

What's changing with this integration

GlobalData's Ava is not a general-purpose chatbot. It's a domain-specific AI trained on decades of industry data, covering sectors from pharmaceuticals to financial services. Until now, customers had to use Ava through GlobalData's own platform or API—a separate window that pulled analysts away from their primary productivity tools. With the Copilot integration, Ava becomes a native component of the Microsoft 365 environment.

From a technical perspective, the integration relies on the Model Context Protocol. MCP, originally spearheaded by Anthropic and now supported by a broad coalition of AI vendors, defines a client-server architecture where AI models access external data via well-defined resources, prompts, and tools. In this case, Microsoft 365 Copilot acts as the MCP client, and GlobalData's Ava server provides the context. This means that when a user types a query in the Copilot chat pane—say, "What's the market forecast for electric vehicles in Europe?"—Copilot routes the request to Ava's MCP server, which retrieves the relevant proprietary data and returns a synthesized answer. The entire interaction adheres to Microsoft's data privacy and compliance boundaries.

GlobalData says the extension will initially support text-based queries, with plans to add structured data retrieval for direct injection into Excel models and PowerPoint slides. The rollout is expected to start with a private preview for select enterprise customers, followed by general availability later in 2026.

What it means for you: practical impact by user type

For business users—analysts, consultants, strategists, and executives—this integration is a time-saving shortcut. Instead of juggling browser tabs or copy-pasting data, you can open the Copilot pane in any Office app and ask Ava for the latest market sizing, competitor benchmarking, or regulatory updates. The AI pulls from GlobalData's research, which includes proprietary databases that are not accessible to public web search. This reduces the risk of relying on outdated or low-quality internet data.

For example, while drafting a report in Word, you might ask: "Ava, give me the top three trends in cloud computing for 2026 with citations." Ava would return a concise summary along with links to full reports. In Excel, you could request historical revenue data for a specific company and have it formatted as a table ready for analysis. In Teams meetings, the Copilot meeting recap could be enriched with real-time market data from Ava when discussing a particular industry.

IT administrators will need to prepare for this extension. Like all Microsoft 365 Copilot extensions, Ava will be manageable via the Integrated Apps portal in the Microsoft 365 admin center. Admins can control which users or groups have access, review the data permissions the extension requires, and set conditional access policies. Because Ava accesses sensitive third-party data, administrators should carefully evaluate how GlobalData's data handling aligns with their organization's compliance requirements. Microsoft typically provides documentation on data flow and residency for Copilot extensions, and GlobalData will likely publish its own technical brief. Look for a "Copilot for Microsoft 365 Extension Planning" guide from Microsoft to understand the security model.

Developers and solution architects can view this as a real-world implementation of MCP. The protocol is still relatively new, and seeing a major enterprise data provider adopt it for Copilot integration validates the standard. If your organization holds proprietary data you'd like to surface in Copilot, you can follow a similar pattern: expose that data through an MCP server, register the extension with Microsoft, and make it available to users. Microsoft's extensibility framework supports both REST API-based integrations and MCP-based ones, but MCP's standardized context exchange may become the preferred method.

Home users or those with personal Microsoft 365 subscriptions won't see Ava—the extension is available only with a GlobalData license, typically an enterprise agreement. So if you're not part of an organization that uses GlobalData, this change doesn't affect your Copilot experience. However, it signals a broader trend: Copilot is becoming a platform for specialized AI tools, and over time, we may see consumer-relevant extensions, such as personal finance or shopping assistants, built on similar principles.

How we got here: the road to Ava in Copilot

The integration didn't happen overnight. Microsoft introduced Copilot for Microsoft 365 in November 2023, and by early 2025 had opened up an extensibility model. Initially, extensions were limited to plugin-like integrations via REST APIs—tools like Jira, Trello, and SAP appeared as early partners. Then, in late 2025, Microsoft announced support for the Model Context Protocol, aligning with the industry's move toward a more efficient, real-time way for AI models to access external resources.

GlobalData has been investing in its AI capabilities since 2022, developing Ava as an internal research tool for its analysts. By 2024, Ava was offered to clients via a web portal and API, allowing them to interact with GlobalData's tagged and structured research data. The integration with Copilot is a logical next step: it puts Ava where clients spend most of their time. According to GlobalData's July 8 announcement, pilot customers have been testing the integration for several months, and the feedback highlighted the productivity gains from having research integrated into the office suite.

This move fits into a larger pattern of "domain AI" assistants. Whereas general-purpose chatbots can answer broad questions, specialized AI tools excel at deep, industry-specific queries. Sales teams have had Copilot for Sales, customer service teams have Copilot for Service, and now knowledge workers in research-heavy roles get Ava. It's a piece of the puzzle that turns Copilot from a universal helper into a tailored workbench.

What to do now: actionable steps

If you're a GlobalData subscriber:

  1. Contact your GlobalData account manager to express interest in the Ava Copilot preview. Early access may be limited, and your organization's feedback could help shape the final product.
  2. Have your IT team review the extension's prerequisites. Microsoft typically requires that your Microsoft 365 environment is on a current channel and that you've configured the necessary admin roles and policies for deploying Copilot extensions.
  3. Plan the deployment: decide which user groups should get access first. Consider a phased rollout starting with power users or teams that heavily rely on research data.
  4. Educate users: once Ava is enabled, provide guidance on how to phrase effective queries. Since Ava's strength lies in its knowledge of GlobalData's taxonomies, users should be trained on how to ask for industry codes, company identifiers, or report names to get accurate results.

If you're not a GlobalData subscriber but interested:

  • Assess whether the integration justifies the cost of a GlobalData license. The value lies in the seamless access to proprietary data within Office. For companies that already spend significant time mining research data, the efficiency gains could be substantial.
  • Keep an eye on other data providers. Providers like Bloomberg, Gartner, and IDC are likely to follow with their own Copilot extensions, so you may see similar integrations for your preferred research source.
  • Evaluate the MCP ecosystem. If your company has proprietary data it wants to surface in Copilot, now is the time to explore MCP server development. Microsoft's documentation and open-source MCP SDKs make it relatively straightforward to create a custom extension.

For everyone else:

  • No immediate action needed, but observe how the Copilot extension marketplace evolves. The availability of specialized AI assistants could change how you evaluate productivity tools in the long run.

Outlook: what to watch next

The Ava integration is one of many expected Copilot extensions in 2026. As the MCP standard matures, we'll likely see a surge of third-party data connectors. Microsoft's Build 2026 conference may reveal more about the Copilot extension roadmap, including deeper integration with Excel's data types and PowerBI. Also watch for how Microsoft handles the complexity of managing multiple extensions within a single corporate tenant—governance, permissions, and data leakage prevention will be critical.

For GlobalData, success with the Copilot integration could push them to develop Ava agents for other platforms, like Google Workspace. The competitive landscape is heating up: Salesforce recently integrated its own Einstein AI into Slack, and Google continues to enhance Duet AI. But by embedding directly into Microsoft 365, Ava gains a foothold in the most widely used productivity suite on the planet.

The message for business technology leaders is clear: the era of siloed AI tools is ending. The assistant that lives inside your document, spreadsheet, or meeting is becoming the new interface for all enterprise data. Whether that's a boon for productivity or a new headache for IT remains to be seen—but for now, the arrival of Ava in Copilot is a concrete step toward that future.