Stack Overflow is reshaping itself into an AI-powered data provider

As part of the Microsoft Ignite conference, Stack Overflow on Tuesday unveiled a new suite of products that it aims to position as a valuable part of an enterprise AI stack. Built on its Stack Internal Enterprise product, this new version of the company looks to reshape its classic problem-solving forum into a tool for translating human expertise into an AI-accessible format.

At its simplest, Stack Internal is an enterprise version of a web forum, but with the additional security and administrative controls you’d expect. The new tools are designed specifically to feed internal AI agents using the Model Context Protocol, with some variations designed specifically for Stack Overflow.

As CEO Prashanth Chandrasekar said, Stack Overflow was already seeing a number of enterprise customers using the API for training, which inspired the new product direction. The company also has content deals with a number of AI labs, allowing it to train models on public Stack Overflow data for a blanket fee.

While Chandrasekar did not name specific clients or personalities, he described the arrangements as “very similar to the Reddit deals”, which brought More than $200 million for that platform.

An important part of the new products is a layer of metadata that Stack Internal releases alongside question-and-answer pairs. This data includes basic information such as who answered the question and when, as well as content scores and more complex assessments of internal coherence. These factors are then used to create an overall reliability score, which tells the AI ​​agent how confident each answer is.

“The customer can set up their own labeling system or we can create that dynamically for them,” said CTO Jody Bailey. “What we will do in the future is leverage this knowledge graph to connect concepts and pieces of information, rather than asking AI systems to do it on their own.”

Although Stack Internal produces tools for enterprise agents, it doesn’t build those agents itself, so it’s difficult to say what the final product will be able to do. But Bailey is particularly excited about the write function, which would allow agents to create their own Stack Overflow queries if they can’t answer a question or notice a knowledge gap.

As Bailey sees it, the read-and-write functionality means that “as we continue to evolve, it will require less and less effort from developers to capture unique information about the way they run their business.”