Work that runs itself.
With an owner behind it.
Flows takes over the work that goes the same way every day or week: matching invoices, routing requests, reordering stock, sending out the figures. It runs on what the organisation has agreed, under a named owner. People keep their time for the work that needs judgement.
The recurring work runs itself. The exception reaches a person.
Automation pays off most where work comes back often and the rules are already known. Three examples of what that can look like.
In practice
Start with one process. Every next one goes faster.
The value is not in one clever automation but in an organisation that can trust it with more and more. That is why it goes in this order.
The growth path shows where an organisation stands and what the logical next step is. To the growth path
Other automation moves data. Flows acts on meaning.
A link between apps knows which field goes where. Flows also knows what that field means, who owns it and what may happen.
| Automation that only moves data | Flows |
|---|---|
| A step knows an app's fields. | A step works with agreed definitions, metrics and data contracts. |
| An AI step runs with the connection's access. | An AI step runs as the flow's owner, with that owner's access and no more. |
| Whoever builds it switches it on. | Someone else approves before a version goes live. |
What, when, where, who and how: decided before it runs.
A flow can only do what an administrator has allowed. That is why an organisation dares to hand it more work.
An AI assistant such as Claude can also draft, test and submit a flow. It never puts one live itself: a reviewer decides.
Every plan has every Flows capability.
No step, control or AI feature sits behind a more expensive plan. What a team builds in its first week works the same way when the whole organisation joins. See the plans
Bring the process that comes back every week.
In one conversation it becomes clear what can run on its own and what a person keeps deciding.






