Before we begin our discussion of Power Virtual Agents Copilot, let’s get you up to speed on Microsoft Copilot in general. In previous posts, we explored Dynamics 365 Copilot and Power Platform Copilot. You don’t need to read those posts before reading this one, but please check them out when you’re done reading about Power Virtual Agents Copilot.
OK, now to the topic at hand… If you’re reading this post about Power Virtual Agents Copilot, we can safely assume that you’re either a consumer of products and services with a deep interest in technology or a leader of a products or services company looking for ways to enhance operational efficiency in customer care.
Either way, you’re likely aware that businesses today are facing a steep rise in customer interactions in the form of inquiries, feedback and complaints. Customers often have questions before making a purchase (e.g., product features, pricing, shipping policies, etc.) and then need support after the purchase is complete. (e.g., help with assembly, missing parts or components, product not working is advertised, etc.) These inquiries can cover a wide range of topics and take a significant amount of time for staff to answer individually. This can lead to long wait times for customers, negatively affecting their overall satisfaction with the organization. Of course, training customer care agents to answer questions efficiently and accurately is also a time-consuming process.
Many organizations have introduced virtual agents to address these ongoing and all-too-common challenges. Microsoft has just introduced a new service to stay ahead of the curve. It’s the Open AI-powered low-code assistant called Power Virtual Agent Copilot, and it’s about to transform customer service delivery as we know it.
Gartner predicts that by 2024 (which is less than six months away, by the way!), 75% of large enterprises will utilize at least four low-code development tools for IT application development and citizen development initiatives. Microsoft’s new copilot offering on virtual agents is a significant step in that pro-dev direction.
With its generative AI feature, Power Virtual Agent Copilot offers structured and natural language processing (NLP) capabilities. You can describe what you want your chatbot — another name for a virtual agent — to do in natural language, and it will generate the code for you. Power Virtual Agent Copilot can help businesses improve customer service delivery by automating tasks and providing 24/7 customer support. It can also help companies to save time and money by reducing customer service staff and repetitive manual processes.
Power Virtual Agent Copilot can simulate conversation with a human user and can be used to answer common customer questions, provide support and even sell products. It uses machine learning to learn from user responses and requests. It identifies the user’s context and triggers the appropriate intent for a reply. This allows the chatbot to generate dynamic and relevant answers for each user.
Below are some of the new features introduced with AI capabilities in Power Virtual Agent:
- Generative answers
Creating “topics” for a chatbot — the traditional approach — can be a difficult and time-consuming task. It requires careful consideration of the purpose of the chatbot, the target audience and the types of conversations that will be had. Additionally, the topics must be relevant to the chatbot’s purpose and be engaging enough to keep users interested. But with Power Virtual Agent Copilot, that is no longer the case. Copilot can automatically create topics for you based on your website or internal knowledge base. This means you can have a chatbot up and running in minutes, with no additional authoring required.
- Enterprise data
The inability to search data from multiple sources can be a significant obstacle for businesses and organizations. Without the ability to search data from multiple sources, gaining a comprehensive understanding of a particular topic or issue can be difficult. For example, suppose a business is trying to understand employees’ preferences. In that case, they may need to search data from multiple sources such as Custom Data, SharePoint, Dataverse, etc. Power Virtual Agent can now connect to multiple enterprise sources — including those just mentioned, plus OneDrive for Business and many customs data sources — allowing chatbots to have rich, multi-turn conversations across multiple systems. It empowers employees to quickly generate content from your enterprise data, allowing you to create relevant and engaging content.
Copilot can be used to:
- Search for information across multiple systems
- Answer questions about the company’s policies and procedures
- Provide summarized information about complex topics
- AI-based chatbot authoring overview
Manual authoring is a complex task requiring much data and training to get it right. The whole dialogue flow knowledge base requires a lot of research and curation to ensure the information is accurate and current. Chatbot’s usefulness can be significantly increased by reducing manual authoring. Power Virtual Agent Copilot can author topics by just describing what you want. Tell it what you want your topic to do, and it will generate a response.
This has been just a first look at the Power Virtual Agent Copilot features recently released by Microsoft. Copilot can help you solve a variety of real-life problems. If you are looking for a way to improve your chatbots’ efficiency, accuracy and scalability, I encourage you to check out this feature. Power Virtual Agent Copilot is still under development, but it has the potential to make chatbot building much easier and faster.
At HCLTech, we empower businesses to grow by using the benefits of AI. Suppose you’re looking for a way to take your customer service to the next level. In that case, Power Virtual Agent Copilot features can help your organization put its IT System on autopilot mode. This means small and large organizations can use Power Virtual Agent to reduce costs and provide round-the-clock customer care to their customers with multilingual support. Note that while I limited the discussion in this post to customer service, there are many other interesting use cases around employee productivity and ecosystem experience. Those explorations are for another post.