A head start.
Then make it yours.
Start with a process, review how it works, and adapt it to your team. Every template becomes a private draft with its own connections and configuration.
Find your starting point.
Explore workflow starting points and the actions your connected services can take.
32 results
Workflows you can use
Each one becomes a private, editable draft in your workspace.
Customer request to coordinated response
An unpublished example: enrich a customer request with Airtable, prepare a response with an agent, route urgency through the Decision Engine, review escalation, then coordinate Slack, Linear and Airtable. Replace example resource identifiers and connect accounts before publishing.
Triage new tickets with AI Decisions
Classify each new ticket, score its severity, and check for escalation. Confident answers label the ticket; unsure ones go to a person.
Route new tickets to the right team
One Route by AI step picks billing, technical or sales and labels the ticket for that team. Anything the AI is unsure about waits for a person.
Send enterprise tickets to their account owner
When a customer from an Enterprise-tier company opens a ticket, assign it to that company's account owner.
Move obvious spam and auto-replies to Spam
Move newsletters, spam and out-of-office replies to the Spam view, but only when the AI is at least 95% sure. Borderline ones get a spam-check label for a person.
Acknowledge and escalate critical requests
For urgent tickets, ask a teammate to acknowledge the request. If it is still unacknowledged after 10 minutes, advance its escalation path. A person records recovery evidence on the ticket.
Flag security reports
Spot reports of vulnerabilities, breaches, exposed data or account takeovers. Confident ones become urgent with a security label; unsure ones go to a person.
Raise priority on urgent words
When the customer's latest message says urgent, asap, outage, down or emergency, raise the priority and label it. No AI needed.
Escalate urgent tickets
When a ticket becomes urgent, move it to the next level of its escalation path.
Escalate frustrated customers
Score how frustrated the customer is from 1 to 5. At 4 or 5, raise the priority and leave a note; the score and confidence are in Decision History.
Post SLA breaches to Slack
When an SLA breaches, post the ticket to a Slack channel so the team can act on it.
Raise priority when an SLA breaches
When an SLA breaches, make the ticket urgent (never lower) and note why.
Post tickets at risk of breaching to Slack
When a ticket's SLA reaches its warn point, post it to a Slack channel before it breaches.
Follow up when the customer goes quiet
When a person on your team replies and the ticket waits on the customer, check in after 3 days, then resolve after 4 more. Any customer reply stops it.
Reopen a ticket when its linked issue is done
When a linked issue, story, pull request or incident is done in its own tool, add a note and move the ticket back to Todo so its owner can tell the customer. Nothing is sent to the customer.
Resolve "thank you" replies
When the customer's latest message is only a thank-you with no new question, resolve the ticket. Only when the AI is at least 95% sure.
Route refund requests to billing
Spot explicit refund requests, label them, raise the priority and leave a note. A person decides the refund; the AI never issues one.
Flag data deletion requests
Spot requests to delete an account or personal data (GDPR and similar). Label, raise the priority and note the legal clock.
Post urgent tickets to a channel
When a ticket becomes urgent, post it to a Slack or Discord channel so the team sees it right away.
Open a Linear issue for bug reports
When the AI is sure a ticket reports a product bug, open a Linear issue and link it on the ticket. Because the AI made the call, a person approves the issue before it is created.
Review urgent tickets
Check a ticket's priority and add an internal review note when it needs attention.
Validate webhook intake
Check who sent a webhook, check its event type, and return a structured result along the selected branch.
Read customer inventory
Read authorized customers from this business and map their identifiers, names, and status.
Read GitHub repositories
Use a named GitHub connection to read the first page of repositories available to that account.
Test ticket routing on a dataset
Run a routing decision on every row of a data table of past messages and the team each belonged to, then see how often it picked the right team.
Look up the customer in your directory
When a ticket arrives, find the customer in your company directory (Active Directory, OpenLDAP) by email, and note who they are for the team.
Publish release notes to Git
A signed webhook with a version and its notes writes them to a file in a Git repository, commits, tags the version and pushes.
Start when a file is added
Runs each time a file finishes uploading to the bucket you choose and reads it, ready for the steps you add next.
Add each new CSV file's rows to a data table
When a CSV file is added to the bucket, read it and insert each row into a data table. Columns are matched by name; other fields are left out.
Summarize each new PDF with AI
When a PDF is added to the bucket, read its text, summarize it with AI and save the summary as a text file under summaries/ in the same bucket.
Open a ticket for each new file
When a file is added to the bucket, open an internal ticket naming the file, its size and type, so the team can pick it up.
Keep a log of every new file
Write one row for each file added to the bucket into the Storage file log data table: its path, size, type, version and checksum. It creates the table the first time.
Turn a workflow into an app people can use.
Start with a NodeHub workflow, shape the interface in Lovable, and keep the support and automation backend here.