What Changes for Enterprise Data Teams
For the last two years, most of the conversation around AI at work has focused on copilots.
Write a prompt. Generate some code. Summarise a document. Speed up a task.
That model is already valuable. But for enterprise data teams, it has also been incomplete.
The real challenge in data is rarely just writing code faster. It is working inside complex, governed environments where pipelines break, schemas change, dashboards need updating, permissions shift, and machine learning workflows need to be monitored long after deployment.
Databricks’ launch of Genie Code reframes AI’s role in that environment: not just as a coding assistant, but as an autonomous agent built to understand, operate, and maintain data systems.
From Suggestions to Execution
What stands out is the shift from AI that supports work to AI that can carry more of it through.
That is the real difference between a copilot and an agent. A copilot waits for instructions. An agent can take a broader objective, break it into steps, and move the work forward with less manual intervention.
- Genie Code: 77.1% Solved tasks
- Leading Coding Agent + Databricks MCP: 32.1% Solved tasks
In practice, that means starting with a goal such as identifying flight delay risks and building a monitoring dashboard, then turning that into a full workflow. From finding the right data and writing the code to delivering the dashboard and supporting it over time.
For enterprise data teams, that shift matters because the challenge has never been code alone. It is context.
Data work depends on more than syntax. It depends on lineage, business meaning, usage patterns, permissions, and governance. That is why this kind of capability becomes more relevant when it is grounded in the platform’s governance layer.
With Unity Catalog at the core, the agent can work within the context of authorised access, organisational metadata, and trusted data assets which is essential if AI is going to move beyond suggestions and into execution.
Three Things That Actually Change
1. The full lifecycle is now in scope
Copilots improved one part of the workflow: writing code faster. Agents expand that support across much more of the data lifecycle. From analysis and model training to deployment, monitoring, and issue resolution.
For enterprise teams, that means less time spent firefighting and more time focused on review, control, and optimisation.
2. Governance is now the foundation, not a checkbox
Governance is becoming a direct enabler of AI adoption.
Teams with strong metadata, lineage, and access controls will be in a much better position to use autonomous agents safely. For teams with fragmented data environments, the level of oversight needed will remain much higher.
In other words, the more mature the governance foundation, the more AI value an organisation can unlock with confidence.
3. Roles shift from execution to judgment
Data scientists still play a key role, but advanced analytics can become more collaborative across teams.
As agent-based workflows mature, the value shifts from writing every line of code to setting clear objectives, reviewing outputs critically, and knowing when human input is needed most.
Genie Code changes the roles inside a data team. Data scientists still provide direction and review, but engineers, analysts, and domain experts can now actively work in notebooks with the assistant and contribute to advanced analytics workflows. It turns data science into a much more collaborative team activity. — Radu Dragusin, Principal Engineer, Danfoss
What to Do Now
Focus first on two or three high-value workflows where the scope is clear and the impact is easy to measure.
Just as importantly, make sure the foundations are in place. The more mature the governance layer, the more confidently AI agents can be used in production. As these systems take on more responsibility, organisations also need a clear way to evaluate output quality and reliability over time.
The copilot era brought AI into the workflow. The next phase is bringing AI into operations. Speak to our Databricks Xperts to explore how to build the right data foundation for governed, production-ready agentic AI.
