An example setup for a customer service team that sorts tickets first thing in the morning: the triage arrives prepared, the people talking to customers are still your team.
It runs on your own server: work files and records are stored on your infrastructure; content processed by AI goes to the model provider you choose, and data sent to external tools goes to the connections you set up.
Chat · support
demo data
Sort the overnight tickets at 8 in the morning and draft a reply to each from the knowledge base. We’ll send them to the customers.
17:40
Question · Ticket triage?
Which rule should refund drafts rely on?
The 14-day refund policy
Leave every refund to the team
Remyral17:41
The workflow draft is ready: refunds and complaints take their own branch, and the urgent category reaches you as a notification.
Morning summary · 08:06Completed
37 tickets sorted
Urgent
3 · flagged
Drafts
34 internal notes
Refund branch
6 tickets
01Workflowfrom the chat or the editor
Morningtickettriage
Built from parts: a schedule pulls new tickets, an AI step with an output schema sorts them and drafts a reply, a switch branches by category.
Workflow · Morning ticket triage
demo data
Morning ticket triageSavedSaveRun
Last run · 08:00 · 37 tickets · 3 urgent
02Setup and flowCustomer support
Setup and flow
A customer service team on a support tool that sorts incoming tickets first thing in the morning.
Setup
Project type
A workspace project with workflow engine 2; a worker that runs HTTP steps.
Trigger
A schedule that pulls new tickets from your support tool’s API, or a signed webhook.
Knowledge base
FAQ and refund policy go into memory as facts or in as a skill.
Network
The credential is in the vault, bound to your support tool’s address only; the project HTTP allow-list holds only that address.
What a day looks like
01
At 08:00 the workflow pulls the tickets that came in overnight from your support tool’s API.
02
For each ticket an AI step returns a result that matches an output schema: category, priority, sentiment and a reply draft based on the knowledge base. “Refund” and “complaint” go down their own branch.
03
The drafts wait in the support tool as internal notes; the team reads, edits and sends them. The urgent category arrives as a notification.
What changes
The morning triage arrives prepared; a per-key concurrency limit keeps one ticket from being handled twice, and you decide how long workflow data is kept (30 days by default).
How it connects
Instead of a ready-made connector, you connect to your support tool’s API with workflows and automations: pull new tickets on a schedule, or start the flow with a signed webhook. Drafts are written back as internal notes through the same API; you set the fields to fit your own setup.
We don’t impose a ready-made flow; design your own, either with AI (describe it in the chat) or yourself (in the workflow editor).
Good to know
The chat panel is for your team; your support tool and your team still talk to the customer, and replies go out with the team’s approval.
For tickets with personal data, you set the privacy notice and retention period to match your data protection policy (KVKK or your local law).
Question card
demo data
A screen from the console · The real console: a question card, where it cannot decide the agent asks and the team chooses.
This example runs on the same infrastructure as all the others: Remyral on your own server, and we build your flow together in the pilot. Just describe the work you want to simplify and the tools you use in your application.
Remyral17:41
