Atlassian continued its recent spate of acquisitions yesterday with the news it has bought Secoda, an AI startup that specializes in semantic understanding of enterprise data. The vendor says this will simplify the process of bringing structured data into its System of Work platform for AI analysis by its Rovo agent. In a Secoda blog post announcing the acquisition, Tiffany To, EVP of Platform and Enterprise at Atlassian, explains:

By bringing Secoda into the Atlassian cloud platform, teams – especially those with complex, enterprise-scale data – will be able to make that leap from data to knowledge, giving employees and agents the context they need to find answers and take action right where they work.

Secoda’s expertise in dynamically learning, organizing and understanding structured data, paired with their AI-driven semantic cataloging will strengthen Rovo’s ability to deliver precise, context-rich insights.

Secoda was founded as a Y Combinator company in 2021 and raised a $14 million series A round last year led by Craft Ventures. Its enterprise customers include 6sense, dialpad, Deezer, Kaufland, LichtBlick and Panasonic, At the time of that fund raise, co-founder Etai Mizrahi explained its mission as piecing together the missing context around siloed enterprise data:

Company data is like a disjointed puzzle at most organizations. Data teams often have large tech stacks, full of applications that don’t communicate with each other, and years of legacy knowledge that is not documented anywhere at all - it lives in people’s heads. Without unified search, every time an employee needs to answer a question related to data, it ends up being a tedious process since there is no centralized source of truth - they are either searching high and low across complicated applications for answers, or tapping the data team on the shoulder and taking their attention away from other tasks.

Essentially, Secoda is a data cataloging tool that allows data teams to curate the knowledge about data that historically has been stored in spreadsheets, documents or just in people’s heads, if at all. It also provides governance and monitoring capabilities. It connects to a wide variety of data sources and BI tools, ranging from Snowflake, BigQuery and MySQL to Tableau, Looker and many others, and can automatically index and categorize data. Where it excels is in then democratizing access to that data for analysis by users across the enterprise, who can simply ask for the data they want in natural language, rather than having to know how to extract it or log a help request with a specialist data team. As Mizrahi explains in his blog post on the acquisition, the goal is:

... giving any team member from data, product development, and business teams the ability to find and use data just as effortlessly as they would browse on the internet.

This latest acquisition follows hard on the heels of Atlassian closing its $610 million purchase of The Browser Company and its $1 billion deal for software engineering analytics platform DX. Terms of this latest transaction, which will have been at a smaller valuation, were not disclosed.

My take

One of the big themes at Atlassian’s recent Team Europe event a couple months ago was the opening up of its longstanding teamwork graph to make it easier to bring in external data. But those external data sources typically don’t use the same data definitions, and so there’s a lot of work required to make sense of their data within the contextual framework of the Atlassian System of Work. You can see how Secoda will transform that effort and make it much simpler to bring those data sources into the same context framework — what I find helpful to think of as an enterprise System of Knowledge. The ability to then surface that knowledge and information within The Browser Company’s Dia browser makes this an even more powerful combination, so it’s an acquisition that seems sure to add a lot of value to the Atlassian platform.

I’m also intrigued to see further evidence of a convergence of not just data but also meaning as vendors start to get their arms around what is needed to really deliver the promise of enterprise AI. A while back, I reflected on the emergence of what I’m calling a Tierless Architecture that transcends traditional application silos and makes any data available for queries or actions by any user or function. At the time, I noted how much work needed to be done to turn all of the data in those separate application silos into more readily transferable information. It’s now becoming clear that generative AI is not only creating much more urgent demand for a truly converged enterprise data layer, but is also enabling a huge leap forward in making it a reality.