Microsoft 365 Copilot Agent Builder Gets a Major Usability Update: What’s Changing in August 2026

Microsoft is making Agent Builder easier to configure with a set of updates designed to simplify how organizations create and manage AI agents. The changes, scheduled to roll out globally in late August 2026, will introduce default capabilities for newly created agents, improve the knowledge configuration experience, and rename the existing “Uploaded Files” option to “Attachments.”

The update is aimed primarily at users who create new agents in Agent Builder and is part of Microsoft’s broader effort to make AI agent development more accessible while still giving organizations control over how those agents operate.

For existing agents, there is an important piece of good news: Microsoft says the update will not affect them. Organizations do not need to rebuild, reconfigure, or otherwise modify agents they have already created.

What is changing in Agent Builder?

The most significant change is that newly created agents will have capabilities enabled by default.

Microsoft says this will provide an experience more closely aligned with Microsoft 365 Copilot, giving new agents access to a broader set of capabilities from the start. The goal is to reduce the amount of initial configuration required when building an agent.

That does not mean organizations are losing control.

The existing capability controls will remain available, allowing creators to customize and scope what an agent can access or do. This is particularly important for organizations building specialized agents for specific departments, workflows, or business processes.

For example, an organization might create one agent designed to help employees find information in internal documentation, while another agent could be built to support a particular business workflow. The ability to adjust capabilities means creators can continue tailoring agents to those different purposes.

In practical terms, the change appears designed to make the default experience more useful without removing the customization options that experienced builders rely on.

A simpler knowledge configuration experience

Another important part of the update involves how creators configure knowledge for their agents.

Knowledge is a central part of an AI agent because it determines what information the agent can use when responding to users or helping with specific tasks. However, configuring knowledge can become complicated, particularly for users who are new to agent development.

Microsoft is therefore introducing an improved configuration experience intended to make knowledge setup easier to understand and manage.

The updated interface should give creators a more intuitive way to work with the information that supports their agents. While the underlying purpose of knowledge configuration remains the same, the interface is being redesigned to make the process clearer.

This could be especially useful as more employees across organizations begin creating their own AI-powered agents rather than relying exclusively on specialist developers or IT teams.

The broader trend is clear: AI tools are increasingly being designed not just for technical users, but for business users who understand a particular process or department and want to turn that knowledge into an AI-assisted workflow.

“Uploaded Files” becomes “Attachments”

There is also a terminology change coming to Agent Builder.

The existing “Uploaded Files” option will be renamed “Attachments.”

According to Microsoft, this is a terminology update rather than a functional change. In other words, organizations should not interpret the new label as a new feature or a change in how files work.

The rename is intended to make the terminology more intuitive and consistent with the updated experience.

For users who already understand the current interface, the change should be relatively straightforward. The important thing to remember is that “Attachments” replaces “Uploaded Files”; the underlying functionality is not being removed or fundamentally changed.

Why the default capabilities matter

At first glance, enabling capabilities by default might sound like a relatively small interface change. For organizations adopting AI agents at scale, however, it could have a meaningful effect on how quickly new agents can be created.

One of the biggest challenges with business AI is finding the right balance between flexibility and control.

If every new agent requires extensive configuration before it can perform useful work, adoption can become slower and more complicated. On the other hand, giving agents too much access without sufficient controls can create governance and security concerns.

Microsoft’s approach with this update attempts to address both sides of that equation.

New agents will start with capabilities enabled, making the initial creation process more straightforward. At the same time, creators can continue to customize and scope capabilities based on their organization’s requirements.

That means the default experience becomes simpler while advanced controls remain available for teams that need them.

For businesses, this could help reduce friction for employees who want to experiment with agents while allowing IT, security, and governance teams to maintain appropriate boundaries.

Existing agents will not be affected

Perhaps the most important operational detail is that existing agents are not being changed by this rollout.

Microsoft states that the update applies to newly created agents. Existing agents will continue to function as they do today.

That distinction should reduce the need for organizations to prepare for a large-scale migration or reconfiguration exercise.

Teams with established agents do not need to rebuild them simply because the Agent Builder experience is changing. Their current capability controls and configurations will remain available.

This is also significant for organizations that have already invested time in designing and testing agents. They can continue operating those agents without being forced onto the new defaults.

When will the update arrive?

The worldwide rollout is scheduled to begin in late August 2026 and is expected to complete by late August 2026.

That means users may see the updated Agent Builder experience during this period as Microsoft makes the changes generally available across supported environments.

As with many cloud service updates, the exact appearance and timing can vary as the rollout progresses. Organizations may therefore see the changes at slightly different times during the rollout window.

The platforms and services affected by the update are Agent Builder and Microsoft 365 Copilot.

What should organizations do?

For most organizations, the answer is simple: nothing is required.

Microsoft has explicitly indicated that there is no required action for existing agents. Current agents will continue to work, and the existing controls for configuring capabilities and knowledge remain available.

However, organizations that regularly create new agents may want to take the opportunity to review their internal guidance.

For example, IT and governance teams could remind agent creators that capabilities are now enabled by default for new agents. Even though creators retain the ability to customize and scope those capabilities, teams may want to establish internal standards for when particular capabilities should be enabled or restricted.

This is not necessarily a response required by Microsoft’s update itself, but it can be a useful governance practice as organizations expand their use of AI agents.

Organizations could also update internal documentation to reflect the terminology change from “Uploaded Files” to “Attachments.” This is a minor adjustment, but keeping training materials and screenshots aligned with the current interface can prevent confusion for employees.

A small update with a broader message

The Agent Builder changes may not represent a dramatic new AI capability, but they reflect an important direction in Microsoft’s AI strategy.

As organizations move from experimenting with individual AI assistants toward deploying agents for real business tasks, ease of configuration becomes increasingly important.

An agent-building platform needs to serve two audiences at the same time. New users want sensible defaults and a straightforward experience. More advanced users need granular controls over capabilities, knowledge, and behavior.

The August 2026 update appears designed around that balance.

Default capabilities can make the first steps easier, while customizable settings preserve the flexibility organizations need. The improved knowledge interface addresses another common source of complexity, and the “Attachments” terminology makes the experience more approachable.

For existing customers, the low-impact rollout is also significant. There is no requirement to modify existing agents, which means businesses can adopt the updated creation experience without disrupting their current deployments.

The bottom line

Microsoft’s Agent Builder update coming in late August 2026 is focused less on introducing a headline-grabbing new AI feature and more on improving the experience of building and configuring agents.

New agents will have capabilities enabled by default, bringing the experience closer to Microsoft 365 Copilot while preserving the ability to customize and scope those capabilities. The knowledge configuration interface will also be updated to make setup more intuitive.

Meanwhile, “Uploaded Files” will simply become “Attachments,” with no functional change associated with the terminology update.

For organizations already using Agent Builder, there is no urgent migration task. Existing agents will remain unaffected.

For teams creating new agents, however, the changes could make the starting experience faster and easier, while still providing the controls needed for specialized business scenarios.

As AI agents become a more common part of workplace technology, improvements like these may prove just as important as new AI models or flashy features. Making agents easier to create, understand, and govern could ultimately be what helps organizations move from AI experimentation to everyday business use.