Databricks Nordic chief says CIOs must lead on AI

Databricks’ Nordic chief, Herman von Greiff, says CIOs need to take a stronger lead in their organizations’ AI work. He joined Databricks in 2020, before ChatGPT made generative AI a concern for company leaders.

Databricks grew from work on Apache Spark and big data. The company has since expanded into a platform for data, analytics, machine learning and AI. Von Greiff says a key change is the convergence of structured and unstructured data. Databricks calls its approach a lakehouse, combining features of data lakes and data warehouses.

Companies increasingly want AI to understand their own information and processes, he says. General language models may understand the internet but lack deep context about a particular business. Organizations that have managed their data systematically have an advantage, though others can catch up by building a sound data foundation.

As data sources, AI models and applications multiply, companies need to track what information they hold, who can use it and how. Databricks’ Unity Catalog is designed to govern data and metadata, including quality and access.

Von Greiff says CIOs also need to control AI spending by matching models to tasks and setting limits on usage. He warns that new AI tools can simply add another layer to legacy systems. Without a usable data foundation, AI agents may struggle to work with company information. He also argues CIOs should coordinate AI efforts across IT and business teams, while involving the business in finding relevant uses.