Automate Shopify work with your agents

Trigger: A customer writes to your helpdesk: the order has not arrived.

With Maplith:

1. Task context: Your agent passes the case to get_task_context and gets the playbook for late orders, if you put one in use, and what its key may do.
2. Route: route from the helpdesk conversation to the Shopify order, by the customer’s email or the order number, with how much Maplith trusts each hop.
3. The map: get_object on the order says where its fulfillment status and shipping label live, by order number; the agent reads both with its own access.
4. The action gate: It drafts the answer, and request_action decides whether a person approves it before it is sent: by your agent in Intercom, only with write access you give it there separately, or by that person in Zendesk.
5. Record and value: record_action records that the answer went out, and how the case ended.

Result: The customer gets an answer drawn from the order itself, approved the way you decided.

Trigger: At month-end, the payments behind a payout and the bank line in your accounts disagree.

With Maplith:

1. Ask: ask: “how do Shopify’s transactions tie to a payout and to the bank?” The answer cites the joins it rests on, and what the map could not establish.
2. Route: route from the transactions to the payout by order and amount, then to bank lines and accounting by payout. The payout itself comes from Shopify Payments or your payment provider, so your agent reads it there with its own access.
3. Findings: Any finding where a scan already saw the two sides disagree, with its evidence and a recommendation.
4. The action gate: request_action for a task for the bookkeeper with the payout, the payments behind it and the bank line they should match, decided by the key’s policy.
5. Record and value: record_action records that the task went to the bookkeeper, and how the check ended.

Result: Your agent knows which keys tie each order to the money in the bank, and which hops to check by hand.

Trigger: At month-end, your bookkeeper asks whether every refund in Shopify was booked against its invoice.

With Maplith:

1. The map: get_object on the transactions says a refund is a transaction of the kind refund, keyed by order number; the agent reads the month’s refunds in Shopify with its own access.
2. Route: route from the refund to its order, then to the invoice in your accounting system by the order reference, with how much Maplith trusts each hop.
3. The action gate: request_action for a task for the bookkeeper with the refund, its order and the reference to book it against, decided by the key’s policy. The month’s refunds go in one task.
4. Record and value: record_action records that the task went to the bookkeeper, and how the check ended.

Result: The month’s refunds reach the bookkeeper as one task, each with its order and the reference to book it against.