Instant cloud flow
What is an instant cloud flow?
An instant cloud flow is a cloud flow that starts when a person, an app, or another flow explicitly asks for it. The start does not come from a system event or a schedule; a button in Power Automate, a control in Power Apps, or a call from a parent flow sets it off.
The trigger can take input, such as a case number, a date, or a choice, and the actions use that input to carry out one bounded task. Instant means the request goes in straight away, not that every result is ready at once: connectors, approvals, and external systems can still take time.
How you start one
The trigger decides who can run the flow and which input fields appear. A manual trigger can ask for text, numbers, dates, choices, and files. A Power Apps trigger receives values from the app. A child flow takes input from its parent and returns results through a response action. After the trigger, an instant flow behaves like any other cloud flow: actions read data, check conditions, and call connectors, and the run history shows what came in and where anything failed.
Ask users only for input they genuinely have to choose. Derive the rest, such as a customer name from a case ID, by reading the source of record again, so the flow never acts on conflicting values from the screen.
When to use an instant cloud flow
A deliberate business action. Let someone submit, release, export, or retry on purpose.
A reusable child flow. Put shared logic behind a clear input and output so other flows can call it.
An action from Power Apps. Run server-side work that does not belong in a single simple data operation.
Do not use a manual button for work that must follow every event reliably; a person can forget to click, or start the same action twice.
Instant, automated or scheduled?
An instant cloud flow fits when a person or a calling app owns the start moment. An automated cloud flow fits when a system event decides, such as a new Dataverse row, and a scheduled cloud flow fits periodic work. So ask who owns the trigger. When a colleague first makes a judgement call, an instant flow makes that decision visible, and validating input while reading critical data back from the source keeps it from acting on stale screen values. For anything reusable or business-critical, place the flow in a solution and use connection references and environment variables so it can move cleanly between development, test, and production.