Use cases
Each recipe below builds one working automation from an empty workflow: which connections it needs, which node to add at each step, what to put in its fields, and what changes once you release it. They are written to be followed in order, and every node they use exists in the editor's Add node palette today.
A workflow is a graph of nodes inside a process. One node is a trigger — a
webhook, a cron schedule, an inbound message, a form submission — and the rest run
after it. Nodes read each other's output with {{ }} expressions over node names, so
{{ Trigger.result.from }} is the trigger's payload as the next node sees it. Nothing
runs in production until you release the workflow and unpause its process; before
that, you test single nodes from the editor.
The recipes
| Recipe | What it does | Connections used |
|---|---|---|
| Route WhatsApp messages to an available team member | Turns every inbound WhatsApp message into a task assigned to whoever on the team is on shift and least loaded, and replies to the customer automatically. | WhatsApp Business |
| Sync form submissions to a Google Sheet | Appends one row per submission of a task's form, updating the existing row when the same person answers twice. | Google Sheets |
| Escalate overdue tasks to Slack | Posts a morning digest of everything past its due date, with buttons that push the due date or close the task from Slack. | Slack |
| Turn an inbound mailbox into tracked work | Polls a support inbox over IMAP, opens a deduplicated task per message and sends the sender an acknowledgement. | Email (IMAP), Email (sending) |
| Triage incoming requests with AI | Classifies whatever an external system posts, routes the urgent ones to a human and files the rest in a database. | AI, Slack, Organization database |
Before you start any of them
- Create the connection first. A node that needs an integration cannot be configured without one — the picker is empty. Every connection is set up in the app's Integrations view; each integration page lists exactly which fields it asks for and how to obtain the credentials.
- Test as you build. Any node can be run on its own from the editor; Honeybase runs the nodes it depends on first. Triggers use their editable sample payload for these test runs, which is why keeping that sample realistic is worth the minute it costs.
- Release when it works. Triggers and schedules only fire from the current released version of an unpaused process. A draft never receives a webhook, an email or a message.
- Automate against the API too. Everything these recipes create — tasks, runs, teams, schedules — is readable and writable through the GraphQL API with an organization API key.