Microsoft Tears Down Its AI Data Wall

Microsoft’s biggest products have generally thrived by letting other companies build on top of them. So Microsoft’s decision to join Apache Ossie, a 1-year-old software industry group aiming to make it easier for AI tools to access data from applications and databases, might not seem like a big deal. But actually it is.

Just four months ago, Microsoft moved to block partners from connecting their data management tools to Microsoft’s Power BI, which companies use to analyze data in charts and graphs. In doing so, Microsoft appeared to be protecting its own data management product, Fabric, at the expense of rival firms such as Databricks and Snowflake.

In other words, Microsoft was opening a new front in the AI data wars, part of traditional enterprise app firms’ efforts to stop customers from using AI from other companies to access data they store in traditional apps.

After our report in May, Microsoft took heat for its decision. We don’t know what’s happened behind the scenes since then, but the company must have had a change of heart. And its decision to join Ossie is in line with CEO Satya Nadella’s strategy of building products that worked those of rivals, such as developing a mobile version of Office for Apple iOS and Google’s Android operating systems over a decade ago.

Salesforce made a similar transition in the past couple of years and now is making it easy for customers to use AI tools such as Claude to access their Salesforce data—for a price.

The Microsoft-Ossie move is important because Snowflake, Salesforce and others have been working since last fall to develop standard definitions of business metrics like gross revenue and net revenue, and having the world’s largest software provider on board could accelerate those efforts. The goal is to avoid situations in which a finance department in a company defines these metrics differently than its sales teams do, which makes it harder for AI to understand what’s going on.

Standardizing such metrics means you can, for instance, use Snowflake’s data management tools to analyze your data in Microsoft Power BI. And when Snowflake customers use AI to analyze this data, it is less expensive to run because it doesn’t need to spend time reasoning to make sense of different metrics as it processes invoices and sets up new employees.

A Microsoft spokesperson confirmed the software giant’s involvement with Apache Ossie but didn’t elaborate on why it changed its mind. The company said in a blog post published overnight that it’s “working with Snowflake and the broader industry” to help “customers define semantic context once and reuse it across ecosystems without duplicating data or business logic.

Microsoft’s about-face is the latest sign that proprietary approaches to data management aren’t feasible, said Chris Lynch, CEO of AtScale, an Apache Ossie member company that sells software for translating raw data into common business metrics. “The future is going to require a heterogeneous stack that lets any database connect to any LLM or SaaS application,” he said.

However, not everything is going as planned for Ossie. Google and Amazon Web Services, which joined an early version of the group last November, are not part of a list of member companies on the Ossie website. Google spokespeople didn’t respond to a request for comment, and AWS spokespeople weren’t immediately available for comment. Google has its own software for defining business metrics, known as LookML, which it gained in its acquisition of data analytics provider Looker in 2020.