Most companies are starting to use more AI agents.
One team may use a coding agent. Another team may use a search agent. Some teams may build their own agents for specific business workflows. Each agent may work well on its own.
The real problem starts when all these agents need to work together.
Today, context often moves manually between tools. Access rules are managed in different places. Governance is added separately for every agent. Over time, this can become difficult to manage.
This is why Databricks Omnigent caught my attention.
Omnigent is an open source meta harness that can sit above different agents and tools. It helps teams connect them, compose them, switch between them, and manage them through a common layer.
It is not trying to replace every agent. It is trying to bring order around them.
This can be especially useful for organizations already using Unity Catalog, AI Gateway, and Databricks workspace identity.
As agent adoption grows, data leaders may soon need to answer questions like:
Which agent is being used?
Which model is behind it?
What data can it access?
What actions can it perform?
How is the usage tracked?
These questions should not be answered across five different vendor consoles.
They need one governed place.
For years, data governance focused mainly on pipelines, tables, dashboards, and user access.
Now governance also needs to cover the agents working on top of that data.
More agents are coming.
The bigger opportunity is making sure they can work together safely, with the right context, access, and control.
That is what makes Omnigent worth watching.
Has anyone started testing Omnigent inside a Databricks workspace? I would be interested to hear your early experience and practical use cases.
