Introducing Flows

You've got your data, your models, and your actions. Flows lets you connect them into a process that runs on its own, without writing code to hold it together.
You build it by dragging blocks onto a canvas and connecting them in order.
How it works
Each block handles one job. Pull data, run it through an LLM, branch on a condition, loop over a list, or call an action that does something. You arrange the steps, and the flow runs them start to finish. When a block isn't enough, you drop in your own code.
What you can build
The shape is always the same. Bring in data, make a decision, take an action. What changes is the context.
Route incoming requests to the right team. Check new records against a rule and flag the exceptions. Summarize documents as they arrive. Pull data on a schedule, run it through a model, and push the result somewhere. Work that used to live in a manual step or a one off script now runs on its own.
Where it gets useful
A Flow doesn't just sit on the canvas. Once you save it, it becomes a real process you can run two ways.
Call it directly from an endpoint, so another system can trigger it. Or hand it to one of your agents as a tool, so the agent runs it mid conversation when it makes sense.
That second part matters. Your agents stop being limited to answering questions. They can run a process you defined, get a result, and act on it.
Available now
Flows is live today, with blocks for data, logic, code, and AI in the first release. Enough to build real processes from day one.
Head to any project, open Flows, and drag a Start block onto the canvas. The full walkthrough is in the changelog.

Still chasing data across tools and teams?
Datazone unifies your data workflows — faster, clearer, smarter.
Get Started – It’s Free