Microsoft is bringing an AI-powered copilot to its business app suite
Photo credit: Jean Luc Ichard / Getty Images
Microsoft today unveiled what it calls the “next generation” of AI product updates for its entire portfolio of business apps. They touch both Power Platform, Microsoft’s set of low-code tools for building apps and workflows, and Dynamics 365, the company’s suite of enterprise resource planning (ERP) and customer relationship management (CRM) tools.
In an interview with TechCrunch, Charles Lamanna, CVP of Business Apps and Platform at Microsoft, described the updates as the logical next step in Microsoft’s automation journey. The new features are based on technology from AI startup OpenAI and were built using the Azure OpenAI service, Microsoft’s service that provides enterprise-specific access to the OpenAI API. The new capabilities follow the introduction of OpenAI text-generating AI models in Power Platform four years ago and more recently Debut of generative AI capabilities in Viva Sales, Microsoft’s seller experience app.
“Over the past four years, we’ve been on a journey to bring generative AI and foundational models to the workplace,” Lamanna said via email, noting that Microsoft has a longstanding partnership with OpenAI to bring the vendor’s technology to Microsoft’s commercialize their own products and through the Azure OpenAI service. “And we’re now at the point where the technology and the product can enable transformative outcomes for customers.”
In Dynamics 365, Microsoft is introducing what it’s calling Copilot (branding GitHub’s Copilot service), which – broadly – aims to automate some of the repetitive sales and customer service tasks.
For example, in Dynamics 365 Sales and Viva Sales, Copilot can help write email responses to customers and create an email summary of a Teams meeting in Outlook. The meeting summary pulls in details from the seller’s CRM, such as: B. Product and pricing information, says Lamanna, and combines it with insights from the recorded Teams call.
“We securely and intelligently access information from customers’ CRM, ERP and other enterprise data sources at runtime,” added Lamanna. “We use large language models to combine the enterprise data with the underlying knowledge to generate answers that are tailored to each customer. Importantly, we don’t use customer data to train the models.”
Over in Dynamics 365 Customer Service, Copilot can compose “contextual responses” to customer inquiries via chat or email, and provide an “interactive chat experience” for customer service agents that draws on knowledge bases and case histories. These complement the new “Conversation Booster” feature in Power Virtual Agents, Microsoft’s chatbot builder, which allows businesses to connect a bot to resources such as a website or knowledge base, and use that data to respond to questions, for which the bot has not been trained.
Conversation Boosters, in turn, complement a new “GPT” model in Microsoft’s AI Builder tool, which companies can use to embed text generation capabilities into their Power Automate and Power Apps solutions. Lamanna says that, for example, a researcher could use it to summarize text from weekly reports and have it sent to his email, while a marketing manager could use the GPT model to create targeted, generated content ideas by typing in specific keywords or topics.
Given Microsoft’s recent foray into generative text – ie Bing Chat – one might hesitate to build an app using the company’s technology lest it go haywire. But Lamanna claims that conversation enhancers and the GPT model — plus Copilot, for that matter — are “grounded in reality” through each client’s CRM, ERP, and other data sources.
“AI-generated content is always clearly identified and users are encouraged to verify accuracy before using it. When relevant, we also cite the sources from which the answer was obtained so that the user can better verify the accuracy of the answer,” Lamanna said. “We have monitoring and controls in place so that we can respond quickly with manual intervention should issues slip through the above lines of defense.”
There is nothing to prevent users from doing not Of course, take the time to verify the accuracy of the content. Time will tell if this becomes a problem; Studies of automation bias, or people’s tendency to place too much trust in AI, suggest that this may be the case.
Fortunately, the rest of Copilot’s features are potentially less problematic.
With Copilot in Dynamics 365 Customer Insights and Dynamics 365 Marketing, marketers can get suggestions about customer segments they may not have considered before and create target segments by describing the segment in their own words. They can also get email campaign ideas by submitting requests to see topics from Copilot, which they generate by pulling from an organization’s existing marketing emails as well as “a range” of Internet sources, it says Lamanna.
Microsoft is catching up in some ways. The CRM elephant in the room, Salesforce, has been adding AI-powered capabilities to its CRM product family for years (or at least trying to). Startups like Glint have also embraced AI, mainly to automate customer service workflows. But as more marketers say they plan to infuse AI into their content strategies, it may not necessarily matter who gets to the point first, but rather who deploys it first on a scale.
“CRM and ERP have long been mission-critical customer and business data sources; However, they often require tedious tasks like manual data entry, content creation, and note-taking,” Lamanna said. “Dynamics 365 Copilot automates these tedious tasks and unleashes the full creativity of employees.”
Beyond the sales realm, Copilot tries to streamline the creation of e-commerce product listings in Dynamics 365 Business Central, Microsoft’s business administration system. According to Lamanna, Copilot can generate product attributes like color, material, and size with descriptions that can be customized by adjusting things like tone of voice, format, and length.
It’s a bit like Shopify’s recently launched AI-generated product description tool, a fact that Lamanna has indirectly acknowledged. He pointed out that Business Central customers using Shopify can publish products with AI-generated descriptions to their Shopify store with “just a few clicks” (hopefully after verifying them for accuracy).
Elsewhere, riding the wave of automation in the supply chain industry, Copilot can proactively flag up issues like weather, finance, and geography in the Microsoft Supply Chain Center that could impact supply chain processes. Supply chain planners can then choose to have Copilot automatically create an email to notify all impacted partners.
Lamanna argues that even simple AI-infused processes like this — automating email — can result in measurable productivity gains.
“According to our recent Business Trends Survey, 9 in 10 workers hope to use AI to reduce repetitive tasks in their job. AI-powered assistants are now tabletops for business apps,” Lamanna said. “We believe Dynamics 365 Copilot will help people get their jobs done faster, so businesses can spend more time on the creative, innovative aspects of their jobs — like building lasting customer relationships.”
As always, the truth lies muddied in some marketing fluff. But what’s clear is that Microsoft isn’t slowing down its investments in AI and automation. As recently as January, Microsoft invested billions more into OpenAI, and the company is excited for a return.
According to Microsoft, Copilot will be added to existing Dynamics 365 licenses such as Dynamics 365 Sales Enterprise and Dynamics 365 Customer Service Enterprise at no additional cost. It will launch in preview from March 6th, with general availability to follow at some point.