Introduction
Microsoft 365 Copilot continues to evolve to compete with solutions like Claude or ChatGPT. While tenant administrators cannot directly intervene in how Copilot processes requests, two recent updates bring substantial improvements to the user experience: conversation preservation (MC1174856) and memory enhancements (MC1158329). These changes, though seemingly minor, significantly strengthen the application's usefulness.
Good to Know
These updates are being progressively rolled out to all commercial and government tenants. Microsoft 365 administrators should be informed to communicate the changes to end users.
Conversation Persistence: Creating Lasting History
How Chat Preservation Works
Conversations with Copilot Chat operate on a model similar to Teams discussions. Each user-Copilot interaction generates a thread of discussion that allows you to:
- View the complete history of past exchanges
- Understand the context and responses provided
- Resume a conversation later without losing previous information
- Refine Copilot's responses based on prior exchanges
This feature proves particularly useful during complex interactions requiring a pause for reflection before resuming the dialogue.
MC1174856: Immediate Conversation Creation
Update MC1174856 (latest status: May 2026, roadmap item 500638) brings a crucial but simple modification: automatic generation of a conversation from the first user message. Previously, you had to wait for several exchanges before a conversation was formally created.
Benefits of this change:
- The conversation entry appears immediately in the navigation pane
- Users can interrupt and resume the dialogue at will
- Context remains intact, facilitating a smooth conversational experience
- Conversation metadata is captured from the start
Attention
This update has experienced significant delays since its originally scheduled date. Deployment to all tenants was scheduled for late June 2026. Check the status via the Microsoft 365 Roadmap dashboard if you don't see this feature yet.
Copilot Memory: Personalizing AI Directives
What is Copilot Memory?
Copilot Memory is a capability introduced about a year ago that allows users to set persistent directives that Copilot should remember to improve its responses. Unlike simple session settings, this memory influences all future interactions with the AI assistant.
Practical Use Cases for Memory
Memory proves particularly effective for:
- Format Preferences: requesting concise or detailed responses based on need
- Technical Examples: specifying the use of Microsoft Graph PowerShell SDK rather than other approaches
- Documentary References: requiring Copilot to cite exclusively Microsoft sources for technology topics
- Search Directives: avoiding information from unreliable or irregular sources
- Communication Style: adapting the tone and structure of responses
MC1158329: Improved Memory Discovery and Management
Update MC1158329 (June 2026) simplifies the interface for managing existing memories. Users can now:
- More easily locate previously created memories
- Modify or delete directives without creating duplicates
- Organize memories in a more intuitive way
How to Create a Copilot Memory
Initiate Creation via Chat
The most direct way to create a memory is to communicate your request directly to Copilot during a conversation. Clearly state the instruction you want Copilot to remember long-term.
Access Chat Settings
Click the menu (three dots) located in the upper right corner of the Copilot Chat interface to open the available options.
Navigate to Personalization
In the dropdown menu, successively select:
- Chat Settings
- Personalization
- Shared memories
Manage Existing Memories
In this section, view, modify, or delete already memorized directives. Each memory appears distinctly with the ability to edit it.
Tip
Don't look for a direct "Add Memory" option in settings. Microsoft deliberately designed the system so that creation happens through natural chat interaction. This allows Copilot to correctly interpret your intention rather than simply filling out a form.
Personalized Nature of Memory
Important to understand: memories are inherently personal. An effective memory for a system administrator may not suit a business analyst. The time investment in defining appropriate directives varies by user:
- Initial Phase: experimentation to identify beneficial directives
- Optimization Phase: gradual refinement based on observed results
- Stabilization Phase: memories naturally adjust to actual needs
Progressive Improvements and User Impact
The Importance of Incremental Updates
While Microsoft 365 Copilot receives criticism regarding its performance compared to competing solutions, progressive improvements like these demonstrate a commitment to continuous optimization.
These updates illustrate a philosophy of gradual improvement:
- Conversational Persistence: resolves major friction in user workflow
- Simplified Memory Management: facilitates the use of existing capabilities
- Staggered Deployment: allows stabilization before generalization
Communication and Awareness
Important
Message Center notifications are often overlooked by IT administrators. Yet, if these updates are not communicated to users, their impact remains minimal. Consider these changes as internal communication topics for teams using Copilot.
To maximize adoption of these features:
- Share updates through your internal communication channels
- Offer training or tutorials on memory creation
- Share examples of effective directives by user role
- Monitor user feedback to identify use cases requiring further optimization
Future Perspectives
These two updates position Microsoft 365 Copilot on a trajectory of continuous improvement. While perfection remains distant for all conversational AI tools, measured but steady steps progressively strengthen the competitiveness of the Microsoft 365 platform.
Conclusion
Automatic conversation preservation and improved memory management constitute concrete advances for the Microsoft 365 Copilot ecosystem. These modifications, though technically simple, address real friction points in the user experience. For organizations, the challenge is to effectively communicate these changes and guide users in optimally exploiting these new capabilities.
