Stack Overflow is turning into an AI data provider

As part of Microsoft’s Ignite conference, Stack Overflow on Tuesday announced a new set of products aimed at positioning them as a valuable part of the enterprise AI stack. Built around the Stack Overflow Internal enterprise product, the company’s new version aims to transform traditional problem-solving forums into tools that translate human expertise into an AI-accessible format.

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

CEO Prashanth Chandrasekar said Stack Overflow had already seen many enterprise customers using its API for training, which inspired the new product direction. The company also has content agreements with a number of AI labs, allowing them to train their models on publicly available Stack Overflow data in exchange for a comprehensive fee.

Chandrasekhar did not name any specific customers or people, but said the deal was “very similar to the Reddit transactions” that brought in more than $200 million to the platform.

A key part of the new product is the layer of metadata that Stack Overflow exports with question and answer pairs. That data includes basic information such as who answered a question and when, as well as more complex assessments of content tags and internal consistency. These factors are then used to create a general confidence score, which tells the AI ​​agent how trustworthy each answer is.

“Customers can set up their own tagging system, or we can create it dynamically,” said CTO Jody Bailey. “What we’re looking to do in the future is actually leverage knowledge graphs to connect concepts and information, rather than requiring AI systems to do it on their own.”

Stack Overflow is creating tools for enterprise agents, but it’s not building those agents themselves, so it’s hard to say what the final product will be able to do. But Bailey is particularly excited about the write feature, which allows agents to create their own Stack Overflow queries if they can’t answer a question or notice a gap in their knowledge.

In Bailey’s view, that read-write capability means that “as we continue to evolve, it will take less and less effort for developers to obtain unique information about how businesses operate.”