Microsoft Ignite is an annual event that showcases the updates and creations that help customers, partners, and developers understand the path that Microsoft is traveling and explore the technology stack's full potential.
This year, Microsoft has announced an array of products and key updates in continuation of new capabilities of Microsoft Copilot Studio and other Copilot + AI stacks. Below is the list of new capabilities that have been introduced newly or updated to the existing functionality of Copilot Studio; also, we shall see what these are in short in this article.
- Faster M365 Extensions
- Generative AI answers improvisation
- Agentic Capabilities
- Integration with Azure AI Studio
- New Analytics
- Security and Governance
- Copilot SDK
- Conversational IVR
Faster M365 Extensions
Microsoft brings the life of the knowledge workers relevantly and easily by providing a lightweight and intuitive tool to develop the Agent builder, which is embedded within the Microsoft 365 Chat interface.
This agent builder connects Microsoft Copilot with the enterprise data and helps workers to quickly refer to the process and support the business-specific scenarios faster.
![Microsoft Copilot]()
Source: Microsoft / microsoft.com
Generative AI answers improvisation
A lot of improvements have been made in the Gen AI sections to improve the quality of answers from the content sources provided. Microsoft is particularly able to achieve this by using enhanced models like GPT 4o / GPT 4o mini to answer the context of the content. Now, even Copilot will be able to reason over the images and the tables present in the source files. If the content source is SharePoint, then Copilot will be able to fetch responses with lower latency and higher quality by improving chunking, document, or passage in the document selection by up to 40 percent.
Also, Microsoft has introduced a Knowledge Curation technique to analyze the knowledge base and provide visibility over which sources are being used and which sections are probably not answered. A guided user experience to filter sources and help users add custom knowledge to fill gaps in the knowledge base.
![Curation technique]()
![Base]()
Source: Microsoft / microsoft.com
Agentic Capabilities
Agents are nothing but autonomous capabilities that can be triggered by events, unlike earlier, which are triggered only by the conversation. Copilot can automate and orchestrate long-running, asynchronous jobs. Humans can be there in the loop at any stage and ask for help. Each interaction can be monitored and understood, which improves the performance.
![Agentic Capabilities]()
Source: Microsoft / microsoft.com
Integration with Azure AI Studio
Connection to Azure AI Search will allow the agents to chat over with the knowledge in Azure AI. Or utilize custom RAG built within Azure AI Studio and supplement it with all the capabilities of the Copilot Studio. Building or Integrating can be easy with the intuitive experience of no code or low code development.
![Azure AI Search]()
Source: Microsoft / microsoft.com
You can also bring your own model. Customers can use models they create or select in Azure AI in their copilots in Copilot Studio. Can enable well-predefined scenarios that rely on custom models within the Copilot Studio topic.
![Azure AI Models]()
Source: Microsoft / microsoft.com
New Analytics
The new and updated analytics experience is highly performant at large scales and provides suggestions on the trends. Provides deeper insights into agent performance from which knowledge source the answers are providing responses and to which answers have poor user satisfaction. KPI information can also show the engagement rate, satisfaction score, and trend comparisons.
![New Analytics]()
Source: Microsoft / microsoft.com
Security & Governance
Microsoft has enabled powerful new capabilities, right from single-tenant app registration, and network isolation to more dominant admin controls. Governing the agentic capabilities with triggers and disabling chat transcript logging proves admin visibility into their authenticated agents.
End users' AI usage can be verified through audit logs in Microsoft Purview.
Here is the list of security and governance controls available for Copilot Studio.
- DLP controls
- Makers audit logs in Microsoft Purview for admins
- Audit logs in Microsoft Sentinel for admins
- Run connectors and flows with end-user credentials
- Sensitivity label for Knowledge with SharePoint
- End-user authentication with certificates
- Maker security warning
- Environment routing
- Maker welcome message
- Autonomous agents governance with DLP
Copilot SDK
The software development kits provide access to endpoints with code that embeds them into applications. These kits can leverage Azure AI Foundry, Semantic Kernel (SK), and other AI services of your choice. The code can extend Copilot Studio conversational agents to multi-agent systems.
![Copilot SDK]()
Source: Microsoft / microsoft.com
Generative IVR
Help build sophisticated IVR systems and engage with the Copilot over the phone. All the Generative AI features of MCS for Voice use cases can be leveraged. One MCS agent can support multiple languages for both chat and voice modalities. The speech can be optimized with custom vocabulary and sensitivity settings. Now it is possible to attach any Dynamics Omnichannel workstream with the IVR.
![Generative IVR]()
Source: Microsoft / microsoft.com