Picking a label
You give it a short list of possible labels, say booking, complaint, or general question. The step picks one from the incoming message and the flow branches on it.
Figuring out a flow and building it yourself takes time. Explain what should happen in plain language and it's ready to go. During a conversation, that same kind of step also recognizes what a message is about, and sends it the right way.
30 days free, no credit card required
Rather have personal help getting started? We'll help you start
Four steps, in the builder you already use.
For example: send a thank you with a link to our Google review once a job is finished.
One sentence summarizing the automation, what starts it, and the steps in order. Not quite right yet? Adjust it before you go on.
The same builder you'd use by hand, already filled in with the steps from the plan.
You look over the concept, ask for a change if you want one, then press the same save button you'd use for a flow you built yourself.
You describe what should happen in your own words. Our AI turns that into a concept: a plan first, then a flow on the canvas once you confirm it. A separate step turns the plan into the actual steps and connections, so the result stays predictable.
That same kind of step can also run while a conversation is happening. It reads the incoming message and picks a label from a list you defined, or pulls data out of it for later in the flow. What happens with that choice is up to the flow you built and turned on yourself.
Type what should happen in the panel next to the builder, the way you'd explain it to a colleague.
You see a summary, the trigger, and the steps. Something off? Adjust it before you continue.
The flow shows up in the same builder you'd use by hand, not saved yet and not switched on.
You press the same save button you'd use for a flow you built yourself. Nothing happens before that click.
This step reads the incoming message and decides where the flow goes next.
You give it a short list of possible labels, say booking, complaint, or general question. The step picks one from the incoming message and the flow branches on it.
An order number, a date, or a name: the step pulls fixed fields from the message and stores them as a variable for later in the flow.
When it doesn't work out, a slow response or an unclear message, the flow continues on the branch you assigned for that. There's no point where a conversation gets stuck on an AI step.
Nothing goes out that you didn't set up yourself for that flow. You see the concept beforehand, or read back afterward why a conversation went a certain way.
Which node types exist, how a plan gets compiled into a flow, and how the AI step behaves on a failure: it's all in the documentation.
30 days free, no credit card required.
Rather have personal help getting started? We'll help you start