Microsoft brings AI-powered Copilot to its business app suite

Image credit: Jean-Luc Ishard/Getty Images

Microsoft today introduced what it calls “next-gen” AI product updates across its business app portfolio. They touch on 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. increase.

In an interview with TechCrunch, Charles Lamanna, CVP of Business Apps and Platforms at Microsoft, described this update as a logical next step in Microsoft’s automation efforts. Leveraging the technology of AI startup OpenAI and built using Azure OpenAI Service, Microsoft’s service that provides enterprise-grade access to OpenAI’s APIs, the new capabilities follow the OpenAI text on Power Platform four years ago. Continue to Generative AI model rollouts. Debut of generative AI capabilities in Viva Sales, Microsoft’s seller experience app.

“Over the past four years, we have been on a journey of bringing generative AI and underlying models into the workplace,” Lamanna said in an email, noting that Microsoft has a long-term partnership with OpenAI and is working with Microsoft’s own products. Commercialize the vendor’s technology, through the Azure OpenAI service. “And we are now at a point where our technology and products can deliver transformative results for our customers.”

With Dynamics 365, Microsoft is launching what they call Copilot (borrowing branding from GitHub’s Copilot service). It’s broadly intended to automate some of the more repetitive sales and customer service tasks.

For example, in Dynamics 365 Sales and Viva Sales, Copilot helps you compose email responses to customers and email summaries of Teams meetings in Outlook. Lamanna said the meeting summary will pull in details from the seller’s CRM, such as product and pricing information, combined with insights from his recorded Teams calls.

“At runtime, we securely and intelligently access information from our customers’ CRM, ERP, and other enterprise data sources,” added Lamanna. “We use language models at scale to combine enterprise data with foundational knowledge to generate tailored responses for each customer. Importantly, we do not use customer data to train the model.”

With Dynamics 365 Customer Service, Copilot can create “contextual answers” ​​to customer queries via chat or email, giving customer service agents an “interactive chat experience” that draws from their knowledge base and case history. increase. They complement the new “conversation booster” feature in Power Virtual Agents, Microsoft’s chatbot builder. This allows the company to connect the bot to resources such as his website or knowledge base and use that data to answer questions the bot has not been trained to answer.

Conversation Booster then complements the new “GPT” model of Microsoft’s AI Builder tool. This allows organizations to embed text generation capabilities into their Power Automate and Power Apps solutions. For example, Lamanna says researchers can use his GPT model to summarize the text of a weekly released report and send it to an email. Marketing managers can also leverage her GPT model to create targeted, generated content ideas by entering specific keywords and topics.

Given Microsoft’s recent forays into generative text (such as Bing Chat), you might be reluctant to develop an app with their technology. But Lamanna claims that Conversation Booster and his GPT model, and his Copilot for that matter, are “realistically grounded” by her CRM, ERP, and other data sources for each customer.

“AI-generated content is always clearly labeled and users are encouraged to verify its accuracy before using it. I will also cite the sources from which the answers were obtained so that they can be better verified,” said Lamanna. “We have monitoring and controls in place to allow for manual intervention and rapid response should an issue slip through the above lines of defense.”

Nothing Hinders the User no Of course, we take the time to check the accuracy of our content. Time will tell if that matters. Research into automation bias, or the tendency of people to put too much trust in AI, suggests that might be the case.

Luckily, the rest of Copilot’s features aren’t too much of a problem.

With Copilot for Dynamics 365 Customer Insights and Dynamics 365 Marketing, marketers receive suggestions about customer segments they may not have previously considered, and help marketers identify target segments by describing them in their own words. You can create You can also enter a request to view topics from his Copilot, which extracts and generates topics from the organization’s existing marketing emails and “various” Internet sources, to get ideas for his email campaigns, says Lamanna.

Microsoft is trying to catch up in some respects. Salesforce, the CRM elephant in the room, has been injecting (or at least trying to inject) his CRM product family with AI-powered features for years. Startups like Glint are also embracing AI, primarily to automate customer service workflows. But with more and more marketers saying they plan to sprinkle AI throughout their content strategy, it doesn’t necessarily matter who delivers the punch first, rather who deploys first. It doesn’t matter if on a large scale.

“CRMs and ERPs have long been mission-critical customer and business data sources, but they often require manual data entry, content generation, note-taking, and other tedious tasks,” says Lamanna. says. “Dynamics 365 Copilot automates these tedious tasks and unlocks the full creativity of our workforce.”

Beyond the sales realm, Copilot in Dynamics 365 Business Central, Microsoft’s business management system, seeks to streamline the creation of e-commerce product listings. According to Lamanna, Copilot can generate product attributes such as color, material, and size, along with descriptions that can be adjusted by adjusting tone of voice, format, length, and more.

It’s a bit like Shopify’s recently introduced AI-generated product description tool, a fact Lamanna indirectly acknowledges. He noted that Business Central customers using Shopify can publish products with AI-generated descriptions to his Shopify store “with just a few clicks” (hopefully with accuracy after confirming).

Elsewhere, riding the wave of automation in the supply chain industry, Copilot at the Microsoft Supply Chain Center proactively flags issues such as weather, finance, and geography that can impact supply chain processes. can do. Supply chain planners can choose to have Copilot automatically draft emails to alert affected partners.

Lamanna argues that even a simple AI-infused process like this (email automation) can lead to measurable productivity gains.

“According to a recent survey of business trends, 9 out of 10 employees would like to use AI to reduce repetitive tasks at work. element,” Lamanna said. “Dynamics 365 Copilot helps employees get work done faster and allows organizations to spend more time on the creative and innovative aspects of work, such as building long-term customer relationships. I believe it can be done.”

As usual, the truth is covered in some marketing fluff. But what is clear is that Microsoft has not slowed its investment in AI and automation. It was just in January that Microsoft invested billions more in OpenAI, and the company was eager for a return on its investment.

According to Microsoft, Copilot is included with existing Dynamics 365 licenses such as Dynamics 365 Sales Enterprise and Dynamics 365 Customer Service Enterprise at no additional cost. It will be available in preview starting March 6, with general availability after that.