Claude · Designed & built · 2026
An agent that takes care of your shipping queues
An agent that works your company's delivery problem queue, matches each ticket to its SOP, checks the order and the parcel's tracking, and leaves a ready-to-review draft written in your own voice. Nothing goes out unless it's preapproved.
The problem
No two shipping tickets are quite alike. A customer asking where their order is, a parcel stuck in customs in Korea, an address that can’t be shipped to, a box that arrived damaged. Each one has its own SOP, and each reply means checking the order, the tracking history and the fulfillment status before writing a word.
What it does
The agent works the queue one SOP view at a time, then sweeps the rest of the shipping inbox. For every ticket it reads the customer’s message, opens the order in Shopify, reads the latest tracking events and the last-mile carrier, and decides which stage of the parcel’s life the ticket belongs to.
Then it fills that SOP’s reply template with the facts and turns it into a reply written in your own voice, translated into the customer’s language when they wrote in one.
Two ways to run it
Draft-only is the default. Every reply waits in the ticket for a person to read, edit and send.
Preapproved send is for the replies a team already trusts. Pick the response types that are safe to go out unread, like a where-is-my-order update with clean tracking, and the agent sends those on its own. Anything outside that list, and anything it flags, still lands as a draft for review.
Routing, not guessing
Every ticket tag maps to one SOP, and the agent opens only the SOP and template that tag points to. Where-is-my-order tickets route before they draft. If tracking shows the parcel delivered, it gets the delivered reply. If it’s sitting in Korean or Taiwanese customs, it goes to the customs SOP. Past 12 business days it goes to Delayed, and a parcel marked lost goes to Lost in Transit.
Every gap gets a reason
Nothing is dropped silently. A ticket the agent can’t handle becomes a flag with the reason, a row in a Notion gap log, and a line in a summary sent to your own Slack at the end of the run. The gap log tells you which SOP needs work next.
Built to stay current
The SOPs live in Notion, where the team edits them. Before every run, the agent checks each local SOP against its Notion page and any SOP that changed since you last verified it runs flag-only until you review it. A changelog records every rule change and the ticket that prompted it, so the agent’s behavior always traces back to a real case.