The new agent upgrades the Databricks Assistant, enabling it to help data practitioners complete repetitive tasks and troubleshoot errors.
Databricks has added a new agent, the Data Science Agent, to the Databricks Assistant, in an effort to help data practitioners automate analytics tasks.
The agent, which is available now in preview and is expected to be rolled out soon to enterprise customers, can be toggled from inside the Assistant window in Notebooks and the SQL Editor, and builds on the Assistant’s functionality to accelerate users’ work, the company said in a blog post.
It said that data practitioners can use the agent for tasks such as data exploration, training machine learning models, and diagnosing and fixing errors.
“You can ask the agent to ‘perform exploratory data analysis on @table to identify interesting patterns’. You can provide additional guidance if you want to focus the exploration on a particular area,” the company said.
For machine learning tasks, the agent can be asked to “train a forecasting model predicting sales in @sales_table”, it added.
With this offering, the company is joining many other analytics software providers in adding agents to help enterprises automate tasks; hyperscalers like Google and Microsoft are integrating similar capabilities into their own data infrastructure services, and Databricks’ biggest rival, Snowflake, is also adding agents to its portfolio of offerings.
Comparing the Databricks Assistant and the Data Science Agent, Forrester VP and principal analyst Charlie Dai said that the new agent upgrades the Assistant from a code-generation copilot to an autonomous agent capable of planning, executing, and iterating on multi-step workflows.
And, noted Samikshya Meher, practice director at Everest Group, the addition of the new agent will cut down the time spent on tedious but necessary steps like data cleaning, model training, and error detection.
“Instead of juggling these repetitive tasks, data practitioners can focus on higher-value analysis… The net efficiency gain comes from both reduced cycle times in development and improved alignment between analytical output and business decision-making needs,” Meher said.
Databricks expects to soon add new capabilities to the Data Science Agent, such as broader context via MCP integration, smarter memory, and faster data discovery, but the company didn’t provide a timeline.
However, it added, “agent mode will grow to orchestrate entire workloads across Databricks. We’re building towards agent workflows for data engineering and beyond.”
To try out the new agent, workspace admins must enable the Assistant agent mode beta from the Databricks preview portal. Once agent mode is enabled, users will be able to toggle the agent from within the Assistant.
