A strong knowledge base is the bedrock of successful Customer Service teams. Investing in it pays off by accelerating future case resolutions. But what if we could reduce the effort of creating these article?
That’s where Copilot’s Customer Knowledge Management Agent can help by providing autonomous AI automation to speed up and simplify knowledge authoring.
Released for Dynamics 365 Customer Service and Contact Center, this feature helps service teams quickly produce knowledge base articles by scanning information in a case and related emails.
The Knowledge Management Agent examines the case content against your existing knowledge base to identify gaps and avoid content duplication. It ensures compliance by removing sensitive data and can be enhanced with custom automated compliance checks tailored to your organisation’s standards.
It is one of several preconfigured agents in Dynamics 365 to help organisations move to AI-first business processes. You can also create custom autonomous agents inside Copilot Studio.
How will this Agent help your Knowledge Base?
The added business value is that it will harvest information across different data sources and compile the knowledge draft in moments to compress the authoring process. Typically, service teams would have to sift through email chains and case notes to begin writing the article when they could solve other cases.
Additionally, actionable insights and tailored analytics empower supervisors to monitor the system’s performance and improve the relevance and quality of knowledge articles, boosting productivity, decision-making accuracy, user satisfaction, and operational efficiency.
Let’s look at an example of how this new AI agent can support knowledge article creation.
It requires opting in and light configuring in the app settings, but you are good to go once you have set it up.
Copilot Drafted Knowledge Articles
When the AI agent is turned on, team member will have the option to propose a new article when a case is closed from the command menu.
If you have an open case, when you click resolve on the command bar, it will present a window with a tick box to “Propose new knowledge article for this case” which will display another window to allow the AI agent to draft the article.
Additionally, clicking the menu item shown in the image below on a closed case will also present the Copilot to enable the AI agent draft an article.
Copilot will automatically begin analysing the resolved case and draft an article that outlines the issue, cause and clear bullet point instructions on how to solve the problem. This is an editable draft and you can ask Copilot to try again with another revision. For example, if it had overlooked a detail, you could ask it in the prompt box to consider adding an extra step.
In the example above, a marketing manager encountered an issue copying email campaigns in their system.
Copilot examined the related communications and case notes to draft an article that outlined the issue as being unable to copy existing email campaigns. It also understood the cause and provided a number of listed resolution steps. A support team person can then follow these instructions in future cases.
As we are happy with its output in our demo, we’ll create the proposal for the article. Of course, we can refine this like any other article and collaborate with other service people and supervisors before it’s published in the knowledge base.
A Couple of Caveats…
Because this update was initially released as a preview, some features were missing, at least on our end. At the time of writing, we didn’t see the option to scan historical cases and publish articles automatically. This will help organisations efficiently fill gaps in their knowledge base, a powerful addition to Customer Service.
In addition, we can’t yet see evidence of knowledge insights. This will enable supervisors and content managers to oversee the Customer Knowledge Management Agent and track the health and utilisation of the knowledge applied by Copilot. We expect to see this option available when it is out of preview.
Copilot’s output will vary depending on your case. More detailed and organised case records with clear information on the issue, cause, and resolution will feed Copilot better data, ultimately delivering a better outcome in your proposed article.
Microsoft updated the general release timeline for the Knowledge Management Agent to October 2025.
Streamlining Knowledge Base Updates with Copilot
Automating the creation of knowledge articles from case data reduces response times and enhances efficiency with shorter knowledge article lifecycles. This adds value by reducing the effort to expand a knowledge base, which improves operational workflows and knowledge management.
If you need help configuring your agent, please contact us today.
