The Story of Contextual AI: Building the Next Generation of Enterprise Language Models

The Story of Contextual AI: Building the Next Generation of Enterprise Language Models

Before language models dominated tech headlines, a team at Facebook AI Research was quietly laying the groundwork for what would become a fundamental shift in AI technology. In a recent episode of Category Visionaries, Douwe Kiela shared how this early work led to the creation of Contextual AI, a company now at the forefront of enterprise AI solutions.

From Research to Reality

The story begins in 2019, when Douwe and his team at Facebook AI Research developed a groundbreaking approach called retrieval augmented generation (RAG). “We wrote the first paper on retrieval augmented generation,” Douwe explains. “And in that paper, we actually showed that what you want to do is train the entire system.”

This research would prove prescient. As language models gained mainstream attention, enterprises discovered that while demos were impressive, production deployments were another matter entirely. The gap between demonstration and deployment became the catalyst for Contextual AI’s formation.

Identifying the Enterprise Challenge

The company was born from a clear observation: enterprises were struggling with AI adoption. “Everybody’s extremely excited about language models. Everybody can see that they’re going to change the world,” Douwe notes. “But at the same time, there’s a lot of frustration, I think, especially in enterprises where you can build very nice demos, but to get these models to actually be production grade, so enterprise grade for a production use case, that requires a lot more work.”

The challenges were numerous: hallucination, attribution issues, data privacy concerns, and prohibitive costs. While others focused on solving these problems individually, Contextual AI saw an opportunity to address them holistically.

Building a Different Kind of AI Company

Instead of chasing artificial general intelligence or consumer applications, Contextual AI focused exclusively on enterprise needs. Their approach leverages open source foundations while adding crucial enterprise capabilities. “We can leverage open source models, which are a very good starting point and kind of contextualize those rather than having to train our entire system at that scale from scratch,” Douwe explains.

This strategy allowed them to focus resources on solving real enterprise problems rather than rebuilding basic capabilities. The company developed a culture centered on learning and collaboration. “We have a very flat hierarchy, and everybody is trying to make this a success,” Douwe shares, contrasting this with larger organizations where “everybody’s kind of trying to promote their own career, maybe sometimes at the expense of others.”

Preparing for the Post-Hype Future

While many AI companies chase the current hype cycle, Contextual AI is building for what comes next. “This hype train is going to stop at some point and so the tide is going to run out and a bunch of people are going to get caught swimming naked,” Douwe predicts. Their focus on specialized solutions and enterprise-grade deployment positions them to survive and thrive beyond the hype.

The Vision: Transforming Work Itself

Looking ahead, Contextual AI’s ambitions extend far beyond improving language models. “What we really want to do is we want to change the way the world works, literally,” Douwe declares. Their goal is to become “the go-to enterprise large language model platform for all enterprises that really care about having high quality language model deployments.”

This vision sees AI not just as a tool, but as a transformation of how work gets done. As Douwe explains, they want to help people become “their own CEOs of their own little groups and teams of coworkers, and being much more effective at doing their job and changing the way the world works.”

In a landscape crowded with AI companies chasing consumer applications and AGI breakthroughs, Contextual AI’s laser focus on enterprise transformation sets them apart. They’re not just building better language models; they’re reshaping how enterprises work, one deployment at a time.