Skip to main content

Microsoft emphasizes AI safety in planning launch of Dynamics 365 Copilot for Business Central

by Jason Gumpert
Editor, MSDynamicsWorld.com

Microsoft has provided Dynamics 365 Business Central with one of the inaugural Copilot capabilities, and product leaders say they have many ideas on where else to take the enterprise AI technology in future product releases. And while interest in the impact of AI tools remains high, Microsoft is also signaling caution in advancing its product roadmap for AI in enterprise business software, including its Dynamics 365 ERP products.

“AI is really going to be a transformational part of how people use [Business Central] going forward,” Microsoft VP for SMB Business Applications Mike Morton told the Directions North America 2023 audience this week. Some of that AI will come in the form of tools specifically for Business Central, but other capabilities will result from connections with Power Platform services, which are adding generative AI capabilities to help speed up the creation of apps, flows, and bots. Still more impact could come from Copilot capabilities in Microsoft 365 applications.

Promoting a safety-first approach

Increasing the reach of AI in the context of protected business data will require vendors like Microsoft to show their commitment to safety and compliance. Morton told the event audience that data privacy will remain a priority for Copilot applications. For example, Microsoft will protect organizational data in the ERP from being used to train the broader large language models (LLMs) used by Azure OpenAI Service, the engine that powers Copilot.

“Your data is your data. We are going to take the data and help fine tune models for your company. What we're not going to do is use it to train the foundational models and be shared across other companies,” Morton said. “And so it will not be the case that when you're typing and you're [using applications], your data can be leaked out, or the way that you communicate can [become] part of some model.”

Morton argued that Microsoft’s commitment to data privacy is strengthened by the fact that customers use such a broad range of Microsoft products and services for their IT needs, from data centers and Azure services to productivity and business solutions. “We're not going to be sending our data to some third party to do these models. Our ability to control that end-to-end ecosystem is going to let us build incredible capabilities and have the privacy and responsible AI to make sure that data stays protected and make sure that the user always remains in control,” he said.

“A human is always in the loop,” Microsoft product manager Mike Borg Cardona explained in another event session, to emphasize that Copilot is not the same as ChatGPT. Copilot uses Azure OpenAI Service and the apps are being designed to have additional safeguards in place to avoid AI hallucinogenic or other disturbing or inappropriate suggestions. Grounding the model ensures the output to the application user is relevant and appropriate by validating against factual data from Business Central, he added. Post-processing of results from the LLM by Copilot verifies formatting, data accuracy related to ERP, and the appropriateness of the suggested content before offering it up for human review.

How else can generative AI improve ERP?

FREE Membership Required to View Full Content:

Joining MSDynamicsWorld.com gives you free, unlimited access to news, analysis, white papers, case studies, product brochures, and more. You can also receive periodic email newsletters with the latest relevant articles and content updates.
Learn more about us here

About Jason Gumpert

As the editor of MSDynamicsWorld.com, Jason oversees all editorial content on the site and at our events, as well as providing site management and strategy. He can be reached at jgumpert@msdynamicsworld.com.

Prior to co-founding MSDynamicsWorld.com, Jason was a Principal Software Consultant at Parametric Technology Corporation (PTC), where he implemented solutions, trained customers, managed software development, and spent some time in the pre-sales engineering organization. He has also held consulting positions at CSC Consulting and Monitor Group.

More about Jason Gumpert