Extend Copilot for Sales with custom data and insights from your own apps

This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Community Hub.

Optimize the sales process by integrating internal and external data with Copilot for Sales. Create enriched email summaries, customized meeting prep documents, and actionable meeting recaps to improve customer interactions. Copilot uses AI to generate insights and suggest actions based on comprehensive data, ensuring sales teams have all the necessary information at their fingertips. Create custom AI experiences using Copilot Studio with zero code. By adding Copilot extensions, users can connect to external data sources and build tailored AI skills.




Eric Boocock, Copilot for Sales Principal Program Manager, shares how Copilot for Sales works and how to build Copilot extensions to connect with other systems for personalized experiences.



Strengthen and inform the sales process.



Bring in data from connected services that provide critical customer intelligence. Check out Copilot for Sales.



Generate personalized meeting prep docs and meeting recaps.



Save time using Copilot for Sales.



Connect data from external systems.



See how to create Copilot extensions with zero code from Copilot Studio using Copilot Actions.



Watch the video here:




00:00 — Copilot for Sales
00:36 — Baseline experiences
02:13 — Bring in external data
03:26 — Use Copilot to respond to emails
03:54 — Meeting prep and meeting recap
04:52 — Security and Responsible AI
06:27 — Copilot extensions
09:09 — Publish Copilot extensions
09:40 — Wrap up


Link References

Check out https://aka.ms/CopilotforSalesExtensibility

Stay updated on monthly releases at https://aka.ms/CopilotforSalesUpdates


Unfamiliar with Microsoft Mechanics?

As Microsoft’s official video series for IT, you can watch and share valuable content and demos of current and upcoming tech from the people who build it at Microsoft.


Keep getting this insider knowledge, join us on social:

Video Transcript:

-What if you could go beyond the insights available in your CRM and collaboration tools to easily bring in data from connected services that provide critical customer intelligence to strengthen and inform the sales process. 


-Today, we’ll look at how this is possible using generative AI with Copilot for Sales, grounded with information from both your internal and external systems. And the zero-code way to light the experience up with Copilot Studio to connect to your data with extensions and build custom AI skills using Copilot actions. 


-Let’s start with your baseline experience using Copilot for Sales, which if you haven’t looked at it in a while, enriches your Copilot for Microsoft 365 experience. For example, email summaries are enriched with data from your CRM. There are suggested CRM updates found in the email to ensure no action is missed and CRM data is current and simple to update from the context of Outlook. 


-Then, to close the loop, suggested email response prompts are geared towards closing a deal that draw on all the context of the email and data from your CRM. Meeting prep documents are more customized with content generated using information from CRM records, along with details from referenced upcoming meetings in Microsoft 365. 


-The brief is formatted to include who the participants are, the opportunity summary, recent meeting insights, email thread summaries, related records from the CRM, and any open customer service cases that the team should be aware of going into our upcoming call. Additionally, Copilot for Sales shares the meeting prep in Teams as team members join meetings. 


-And meeting recaps are made more relatable and actionable with the rich AI capabilities of Copilot for Microsoft 365, used to summarize the meeting and Copilot for Sales, bringing in sales-specific actions like follow-up tasks to create in CRM, as well as a record of questions asked, conversational KPIs, and keywords mentioned. In these examples, we ground your generative AI experience with information from your CRM and Microsoft 365. Next, let me show you how your experience is enriched further by bringing in data from external connected systems and tailored custom skills. 


-As I open an email from a prospective customer, Alberto, in addition to what we just saw, we’ve created an extension, Contoso Customer Intelligence, which surfaces additional details about the customer, including other people at Contoso he’s worked with in the past, as well as product pages he’s favorited while signed into our company website. And in a moment, I’ll show you how we enabled this in Copilot Studio. 


-Within the Copilot for Sales sidecar, the custom extension Contoso Customer Intelligence displays expanded company insights and suggested actions to update opportunity details and schedule a meeting. Given the prior work history with Alberto’s company, Fourth Coffee, I can use the Copilot chat experience within Outlook to then prompt for past opportunities and agreements with Fourth Coffee. And I can find the details about Fourth Coffee’s previous purchases with us. 


-These results also include information from our Contoso Customer Intelligence connected app. With all this context, I can respond to Alberto’s email mentioning those additional details. I’ve used Copilot to start an email here, along with an invitation to a meeting to discuss the espresso machines that he’s interested in. I’ve also been able to proactively include a 5% discount opportunity if Alberto completes his purchase this month. 


-So, as I’ve shown, generative AI can now gain context from available information across my CRM, Microsoft 365, and Connected Systems, to inform and improve my customer interactions. And there’s even more that Copilot for Sales can do to help me save time with additional insights. With Copilot in Word, I can generate personalized meeting prep documents connected to my CRM and external applications through Copilot for Sales. In my case, I’ll use Copilot with a forward slash, to reference an upcoming sales meeting called Follow-Up: Fourth Coffee. 


-This generated document includes details about the meeting, participant information, topics for the discussion, and even account-specific information extracted from CRM and our connected external app, as well as open tasks all within seconds. Also, moving ahead to the day of the meeting, I can use the integrated meeting recap to take care of follow-up actions. 


-Here, the after-call summary captures notes including integrated follow-up tasks with our Copilot extension. That’s one example of a connected and integrated Copilot for Sales experience. Now, since this is Mechanics, let me explain how the different surfaces were connected, enabling Copilot to securely access the information it needed to inform responses. 


-The Copilot for Sales system is comprised of large language models, along with the Microsoft Graph, with user-level activities and data, and Microsoft 365 app experiences, and it connects to your sales data in CRM systems from either Salesforce or Dynamics 365, as well as connected external systems using Copilot extensions. With the right grounding data, including external data that’s accessible, here’s how responses are then generated. Beginning with the prompt, Copilot utilizes Retrieval Augmented Generation to orchestrate information search from the Microsoft Graph, CRM system, and connected external systems. 


-Copilot for Sales also incorporates AI-generated skills to guide it through multi-step tasks like proposal creation, where it gathers data from diverse sources using its orchestrator and predefined logic. The gathered information is then presented to the language model, temporarily enriching its knowledge to generate an informed response. Note that the contextual information provided to the LLM remains separate and is never used in its training. 


-And, of course, Copilot for Sales follows Microsoft’s documented practices for responsible AI. All services are designed to meet regulations required in your country, region, and industry, in the same way that other Microsoft Cloud services do. When you use Copilot for Sales, it includes integrated scenarios with Microsoft 365 and your CRM out of the box. 


-That said, to connect to data from external systems, you’ll use Copilot extensions. Let me first show you how you’d create one with zero code from Copilot Studio, using Copilot actions to build a custom AI experience using your external data. Now, remember the email summary we showed you before? I’m going to show you how we extended that experience. For that, I need to head over to Copilot Studio to hook into our customer intelligence application. To get to Copilot Studio, you can navigate to copilotstudio.microsoft.com. 


-Next, I’ll expand the left menu and move to Copilot for Sales. I’ll add an action. From there, I’ll select my connector, which is a Copilot extension type for my customer intelligence application. I’ll paste in the action name and description. In this case, the extension exposes related marketing data as well as public data, including news and events associated to key customers. This description is important because Copilot for Sales will use it to figure out when to invoke this action. 


-Next, I’ll choose from the connector’s available actions to pull the right insights from our email summary with get engagement insights. I can add an action name, then a description for what it does. Again, this description is important because it is used by the AI as the instructions for what to do. 


-Here, we are saying to look into our Contoso Customer Intelligence data set for insights related to an email in Outlook. To save some time, I’ll skip a few of the options and optional steps here. I can repeat this process if I want to add additional actions. You can then optionally add conversation starters, which are suggested prompts, but in this case, to save a little time, I’ll skip this. 


-From here, I can test it out to make sure everything is working as expected. I’ll enter a test prompt, and make sure it responds as expected, and even dig into more details about the logic and information connected with our extension. Everything looks good. And optionally, I can also test this out in microsoft365.com or Outlook. Now, I can finalize and publish it. You can do that from the add connector action experience in the manage tab in Copilot Studio. 


-I’ll continue in the wizard, and to save time, I’ve added the icon images and background color already, as well as who has permissions to use it. So I just need to hit skip and publish, and everything is published in Copilot for Sales. And there’s one more thing to do to get it working in Copilot for Microsoft 365. From the new actions tab, I’ll hit publish, and then I’ll choose Copilot for Microsoft 365. And from there, I can publish it so that it’s available in both Copilot experiences. 


-And by the way, if you’re an ISV or partner building Copilot actions, the same process applies. Plus, you can get your solutions certified using the Microsoft Partner Center, and make them discoverable by organizations in the Microsoft 365 admin center. So that was a quick tour of how Copilot for Sales can optimize the sales process using generative AI, the mechanics of how it works, and how to build Copilot extensions to connect with other systems for even more personalized experiences. 


-For more information, check out aka.ms/CopilotforSalesExtensibility. And to stay up to date on monthly releases, check out aka.ms/CopilotforSalesUpdates. Keep watching Microsoft Mechanics for all the latest AI updates, and thanks for watching.

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.