Microsoft CEO Satya Nadella. Microsoft's long-standing recipe for success with its core products has been to let other companies build add-ons on top of them. So the company's decision to join Apache Ossie, a software industry group that is barely a year old, may seem unremarkable at first glance. The group aims to make it easier for AI tools to access data across various applications and databases, but in reality the move is significant.

Just four months earlier, Microsoft had imposed restrictions preventing rival data management tools from connecting to Microsoft Power BI, the product businesses use to turn data into charts for analysis. At the time, Microsoft appeared to be protecting its own data management product, Fabric, at the expense of competitors such as Databricks and Snowflake. In other words, Microsoft had opened a new front in the battle over AI data. This is also the habitual thinking of traditional enterprise software vendors: stop customers from using other companies' AI tools to read data stored inside their own legacy applications.

After we published our report on the matter in May, Microsoft's policy drew heavy criticism. We don't know what happened behind the scenes afterward, but the company clearly changed its mind. Microsoft's decision to join Ossie also fits CEO Satya Nadella's consistent strategic approach: making its own products compatible with rivals' products, just as more than a decade ago Microsoft built mobile versions of Office specifically for Apple's iOS and Google's Android. Over the past few years, Salesforce has completed a similar pivot. Now customers can pay to use external AI tools such as Claude to access data stored inside Salesforce.

Microsoft's entry into Ossie is significant. Since last autumn, companies such as Snowflake and Salesforce have been pushing to standardize the definitions of business metrics like total revenue and net revenue. The participation of Microsoft, the world's largest software vendor, is expected to accelerate that work. The goal of standardization is to resolve conflicts over metric definitions inside enterprises 鈥?for example, when finance and sales departments define the same business metric differently, which makes it hard for AI to understand business data. Once metrics are standardized, a user could, for instance, process data with Snowflake's data management tools and then analyze it in Microsoft Power BI. Snowflake customers would also see lower running costs when using AI to analyze that data: when AI handles tasks such as processing invoices or creating new employee accounts, it no longer has to burn computing power interpreting a patchwork of custom business metrics.

A Microsoft spokesperson confirmed that the software giant has joined the Apache Ossie project but did not explain in detail the reasons for the policy shift. In a blog post published overnight, Microsoft said: "We are working with Snowflake and a broader set of industry partners to help customers define semantic business context once and reuse it across multiple technology ecosystems, without repeatedly copying data or rebuilding business logic."

Chris Lynch, CEO of AtScale, a member company of Apache Ossie, said Microsoft's sharp reversal once again proves that a closed, proprietary approach to data management no longer works. AtScale's main software products convert raw data into unified business metrics. He said: "The future is definitely a heterogeneous technology stack, where any database can connect to any large language model and any SaaS application."

Still, the Ossie project has not been all smooth sailing. Google and Amazon Web Services (AWS), both of which praised the project last November, do not currently appear on the member list on Ossie's official website. A Google spokesperson did not respond to a request for comment, and an AWS spokesperson could not be reached. Google has its own business metric definition tool, LookML, which it acquired through its 2020 purchase of data analytics vendor Looker.