Hey everyone, this is Venice from Rockhop, and in this video we are talking about Copilot. There is so much buzz right now about Microsoft’s new Copilot releases, and this video breaks down what is going on with Copilot, the differences between these different Copilots because Microsoft has named everything Copilot, what needs to be known about Copilot, how to prepare, and how to start using Copilot. All of these things are discussed in this video.
First off, it is important to break down what is going on with Copilot and the differences in references. Microsoft has named anything that utilizes generative AI in their products as Copilot. There are different Copilots — there is a Copilot in Power Apps, in Power BI. There will be Copilot 365, which will be in all of the different 365 products, such as Copilot Outlook (though it has not been commonly called that), and so forth. There is Copilot Edge, which will be demonstrated, and Copilot Studio, which will be broken down.
This video starts in Copilot Edge, a practical and actionable way to get started utilizing Copilot today. Then Copilot Studio will be shown, illustrating how to build custom Copilots, followed by a discussion of Copilot 365 and what is to come.
The first example is utilizing Copilot in Edge or Bing Chat. Initially, there is a question of the difference between using Bing Chat versus ChatGPT, as they seem similar. The advantage of using Copilot in Edge is the ability to feed Copilot data, allowing it to answer questions within the context of the Edge browser. This elevates generative AI beyond just using the internet. Copilot can answer specific questions about documents or videos being worked on, accelerating work.
Within the context of Edge, Copilot understands the current page being viewed. This allows for targeted assistance, such as summarizing content, without consuming the entire source. For example, Copilot can summarize ingredients for a recipe from a video page. While this requires feeding Copilot the source content, once done, Copilot provides generative AI that understands the page context. This is a valuable starting point for integrating AI into research and work.
Next is Copilot Studio — an evolution of Power Virtual Agents. Copilot Studio allows for building custom chatbots that incorporate generative AI. This process can start by selecting a website as a data source. In the example shown, the New York Times Cooking website is supplied, enabling the chatbot to respond using generative AI based on that source.
Copilot Studio still leverages Power Automate, enabling access to other data sources. Options include adding websites, internal SharePoint sites, or up to 100 files under a specific size. This approach combines Copilot with context and data to create richer interactions.
Copilot 365 represents the integration of generative AI across all Microsoft 365 products, including Teams, Word, Outlook, PowerPoint, and Excel. Copilot 365 carries the context of the user’s tenant, unlike previous examples where context needed to be supplied manually. This enables Copilot to work seamlessly with emails, documents, and other tenant data.
As organizations adopt Copilot, plugins will become increasingly important. Plugins enable access to enterprise data outside the Microsoft 365 tenant, authenticated via APIs, flows, or connectors. This will be critical for making external business data actionable in Copilot.
Other considerations include governance within Power Platform environments to protect enterprise data and ensure security, even as users create their own Copilots. Some scenarios may involve on-premises data sources without APIs, making Microsoft Fabric a potential solution to unify disparate data sources.
Ultimately, success with Copilot requires readiness — data readiness, governance readiness, connector readiness, and user adoption readiness. Organizations should encourage the use of Copilot and brainstorm generative AI use cases to maximize adoption and results.