With the preview of Copilot in Dynamics 365 Customer Service, it’s easy to bring next-generation AI to your customer support organization. Enable Copilot so your agents can start benefiting from these capabilities today:
- Chat with Copilot. Agents can diagnose customer problems and find solutions with Copilot-assisted search of internal knowledge bases.
- Generate live conversation responses. Built-in generative AI suggests appropriate responses faster than agents can type. Agents can handle more cases in less time regardless of how a customer reaches out for support.
- Send expertly crafted emails. Agents can use predefined and custom prompts to create relevant and personalized email responses in seconds.
- View the Copilot case summary. Agents can view a summary of case highlights on the default case forms, shaving minutes off their case ramp-up times.
Before you begin your transformation journey, we recommend learning more about how Copilot in Dynamics 365 and Power Platform delivers enterprise-ready AI. This will help answer questions about how Copilot works and how it keeps your business data secure.
When you’re ready to enable Copilot, follow these steps to turn your agents into super-agents in just 30 days:
- Identify the support cases and agent groups that will benefit most from Copilot.
- Understand how your organization is storing data.
- Examine your customizations and eliminate potential blockers.
Identify the support cases and agent groups that will benefit most from copilot
The best way to start using Copilot in Dynamics 365 Customer Service is by identifying a high-value use case where agents need to search for information and draft responses to customers. We recommend starting with the agents who have to look up answers in organizational knowledge bases.
You can roll out Copilot in phases using the App Profile Manager in the Customer Service admin center. Whether you’re using out-of-the-box or custom profiles for your agents, make sure you have a phased plan for adding Copilot to each agent profile.
Understand how your organization is storing data
Copilot is built on top of Dynamics 365 Customer Service cases and knowledge management. If your organization has content outside of the Dynamics 365 Customer Service application that support agents need, the best way to make it available to Copilot is to sync it with Dataverse using integrated search.
Copilot pre-processes your knowledge articles weekly and generates snippets of documents that it can easily discover. When an agent asks Copilot a question, it retrieves the most relevant snippets and provides those to the AI to generate a response.
The primary source of knowledge is your organization’s data. Copilot currently processes 100,000 of your latest published knowledge articles. It’s crucial that these are high-quality sources that your agents would normally reference when they’re helping customers. Be sure your implementation plan includes revisiting your published knowledge content. If you wouldn’t want an agent to use the source document itself, don’t expose that same document to Copilot.
While Copilot can understand your cases and structured knowledge base today, it’s rapidly expanding into more key agent workflows across the customer journey. Over time, Copilot will have visibility into all the unstructured data in your support organization, like case history, conversations, and notes. As with your knowledge articles, high-quality historical data will produce better Copilot results for your agents.
Examine your customizations and eliminate potential blockers
Copilot appears in the productivity pane in the Customer Service workspace application. It’s important to verify whether you have customizations that may interfere with Copilot. For example, a problem could occur if the app launches a particular productivity pane tool by default every time a case loads. Just-in-time access to data and other customizations can also affect Copilot availability.
Additionally, Copilot considers role-based security access when it’s searching your knowledge content. If an agent doesn’t have access to certain articles, the content of those articles won’t be in Copilot’s responses to that agent’s queries.
Copilot AI capabilities for Customer Service are also available in single-session and custom applications that are built on Power Platform. In Power Apps Studio, you can configure Copilot features for custom case forms and custom apps.
Incorporate agent feedback and insights for continuous improvement
Every conversation that a support agent has with Copilot is stored in Dataverse as an activity and linked to the agent’s cases. If agents provide feedback on a Copilot response, their feedback is in Dataverse, too. This enables agents to play a critical role in supporting the organization’s feedback loop and makes Copilot interactions available to review on an ongoing basis.
It’s easy to export that data for analysis. You can track your most important contact center metrics like average handle time, customer satisfaction scores, and throughput trends when Copilot is in use. You can also build processes to review responses that agents have marked with a thumbs-down. Both are important steps in understanding Copilot usage in your organization and finding opportunities to improve your content.
Plan for future updates
We plan to update Copilot experiences monthly with features and enhancements. The AI models and prompts will be updated regularly behind the scenes, and you’ll see the quality of Copilot’s responses improve over time. The following features are already on our roadmap:
- Conversation summary with Copilot. We’re enhancing the conversation summary capability to enable seamless conversation handoffs and wrap-ups.
- Out-of-box analytics. These reports will give supervisors insight into how their agents are using Copilot and its impact on their tracked support metrics.
- SharePoint support for Copilot knowledge. Organizations who have their support content outside of Dynamics today can easily connect to a SharePoint content source.
AI solutions built responsibly
Enterprise grade data privacy at its core. Azure OpenAI offers a range of privacy features, including data encryption and secure storage. It allows users to control access to their data and provides detailed auditing and monitoring capabilities. Copilot is built on Azure OpenAI, so enterprises can rest assured that it offers the same level of data privacy and protection.
Responsible AI by design. We are committed to creating responsible AI by design. Our work is guided by a core set of principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. We are putting those principles into practice across the company to develop and deploy AI that will have a positive impact on society.
Learn more about copilot AI
Watch a video to learn how copilot AI searches company knowledge sources and generates optimized responses in a single click.
Read the documentation to learn how to configure Copilot features in Customer Service.