Handle "Where Is My Order?" Right Inside Slack

That round-trip is small and constant, which is the worst kind of cost. This guide shows how to kill it: a Macha agent that lives in Slack, answers an @macha mention about any order, and replies in-thread with the Shopify status, tracking, and delivery estimate — no tab-switching, no shared admin logins, no waiting on the one person who "knows where to look."

Macha is an AI agent layer that sits on top of the helpdesk and tools you already run. For this use case the relevant connectors are Slack and Shopify; you don't replace anything, you just give your team a faster way to ask.

What the agent actually does

The workflow is deliberately small — two tools, one trigger:

A real exchange looks like this:

Sarah: @macha what's the status of order #2041? Macha: Order #2041 for [email protected]: • 2x Blue Widget ($29.99 each) • Shipped via USPS — tracking #9400111 • Estimated delivery: March 7

No form, no slash-command syntax to memorize, no leaving the thread the question was already being discussed in. The agent reads the mention, extracts the order number, calls Shopify's Get Order, and formats the reply. Because Get Order returns full line items, fulfillment status, and payment status — not just a tracking string — the same agent can answer the follow-ups too: "was that paid?", "what was in it?", "which carrier?"

Why in-Slack, internal lookup is its own thing

When you search for WISMO tooling, nearly everything is a customer-facing storefront widget — Zipchat, eesel, Frontdesk, ResponseCX, and the rest pull Shopify tracking to deflect tickets before they reach a human, and some report 35–50% deflection on order-status questions (Zipchat). That's a real and separate job.

Internal, team-facing lookup is thinner ground. The options that exist are mostly slash-command order search (EZ Notify), Shopify's own internal "Spy" bot, or order notification apps (Reporty, Sales Bot, Ting) that push alerts but can't answer an ad-hoc question. Macha's angle is that the same agent layer answers in conversation — @mention, plain English — and the very same Shopify + helpdesk connection can power your customer-facing replies too. You build the order-lookup capability once and point it at whichever surface needs it.

Build it in Macha, step by step

You can do this two ways: describe it in plain language and let Macha's AI builder assemble the agent, or wire it by hand. Both end in the same place.

1. Connect Slack and Shopify

Add the two connectors from the dashboard. Slack uses OAuth (you install the Macha bot into your workspace and pick the channels it can see); Shopify connects with an API key and exposes order, product, and customer lookups.

2. Set the trigger to a Slack App Mention

Create the agent and give it the App Mention trigger so it only acts when explicitly summoned — not on every message in a busy channel.

One detail worth knowing: Slack triggers carry a debounce setting (Slack DMs default to ~10 seconds, messaging events to ~15) so a flurry of edits or a multi-line message resolves into a single clean run instead of firing repeatedly. You can adjust it on the trigger's debounce chip. For an order-lookup bot the default is fine.

3. Give it the Shopify and Slack tools

Add Shopify → Get Order and Slack → Reply in Thread. If you want the agent to handle vaguer asks like "where's the order for [email protected]?", add Search Orders too (it matches by customer email, fulfillment status, financial status, or date range) so the agent can find the order before fetching its detail.

4. Write the instructions

Keep them tight. Something like:

You are an internal order-lookup assistant for the support team. When a teammate @mentions you asking about an order, identify the order number (formats like #2041 or 2041) or the customer email. Use Get Order — or Search Orders if only an email is given — then reply in the thread with: customer email, line items with quantities, fulfillment/shipping status, carrier and tracking number, and estimated delivery. If you can't find the order, say so plainly and ask for the order number or email. Never invent a tracking number.

That last line matters: an order-status agent should refuse to guess. If Shopify returns nothing, the right answer is "I couldn't find #2041 — can you confirm the number?", not a fabricated ETA.

5. Pick a model and test

Order lookup is a light reasoning task — read a mention, pull a record, format a reply — so a small, cheap model handles it well. In Macha each model carries a per-action credit cost; the default GPT-5.4 Mini is 1 credit per action, with the range running roughly 0.5–9 credits depending on the model you choose. A single lookup is usually two actions (the Shopify call and the Slack reply), so a few credits per question. See pricing for how credits map to plans.

Run a test mention before you turn it loose: @macha status of #2041? in a private channel, confirm the reply format, then invite the bot into #cs or wherever the questions actually land.

Where this fits in a bigger setup

The order-lookup agent is a building block, not the whole picture. A few natural extensions:

• Same capability, customer-facing. Point the same Shopify Get Order tool at a customer-facing agent on your helpdesk so that the public "where is my order?" tickets get auto-resolved with the same order data — internal and external WISMO from one connection.
• Escalation in reverse. Have a helpdesk agent relay urgent or VIP order tickets into Slack so the team sees them where they already work.
• Sub-agents for safety. If you wrap order lookup as a sub-agent, note that sub-agents are read-only in interactive chat and only gain write permissions when running autonomously via a trigger — a useful guardrail when a lookup helper shouldn't be issuing refunds.

Watch-outs and when not to use it

This is a genuinely low-risk automation — it reads order data and posts a message; it doesn't refund, cancel, or email customers — but a few honest caveats:

• It's only as fresh as Shopify. Tracking and "estimated delivery" reflect what Shopify holds. If your carrier or 3PL updates lag, the agent will faithfully report stale data. For true real-time carrier events, you'd layer in a tracking provider, not just Shopify.
• Channel scope and permissions. The Macha bot can read the channels you add it to. Don't drop an order-lookup bot — which can surface customer emails and addresses — into a wide-open company channel. Keep it in support/ops channels and treat order data as the customer PII that it is.
• One order at a time. This pattern answers specific, identified orders. It is not a reporting tool — "how many orders shipped today?" is a different job (a scheduled report or a Study), not an @mention.
• You may not need a bot at all. If order questions are rare and one person owns Shopify admin comfortably, a Slack bot is overhead. The value shows up when lookups are frequent, spread across people, and interrupting someone who shouldn't have to context-switch.

If your actual bottleneck is customers asking where their order is, build the customer-facing version on your helpdesk instead — the internal Slack bot is for your team.

FAQ

Does this replace a customer-facing WISMO chatbot? No — it's the team-facing complement. This agent answers your staff in Slack. The same Shopify connection can power a separate customer-facing agent on your helpdesk; you're not choosing one or the other.

Can the agent look up an order by customer email instead of order number? Yes. Add Shopify's Search Orders tool, which matches by email, fulfillment status, financial status, or date range. The agent finds the order, then calls Get Order for the full detail.

Does Macha replace Shopify or my helpdesk? No. Macha is an AI agent layer on top of the tools you already use. It reads from Shopify and acts in Slack through their official connectors — nothing gets migrated or replaced.

How much does each lookup cost? Macha bills in credits per AI action. The default model (GPT-5.4 Mini) is 1 credit per action, ranging ~0.5–9 by model. A typical lookup is a couple of actions. See pricing.

Is it secure to expose order data in Slack? Order data includes customer PII, so scope the bot to internal support/ops channels rather than open company-wide channels, and add it only to channels where that data is appropriate.

Can I build it without configuring everything by hand? Yes. Describe the workflow in plain language and Macha's AI builder assembles the agent — trigger, tools, instructions, and model — for you to review and tweak.

Want to try it? Connect Slack and Shopify, describe the order-lookup agent, and test it with one @mention. Macha offers a 7-day free trial, no credit card required — start in the docs or see pricing.