Sponsor slot Your tool here: reach small-business owners automating with AI · $300/week ›

Categories / Operations & Inventory

BOM-checked purchase orders

Organic sports drink brand · back live in two hours after a $3,000 ordering mistake

CostFreemium: Katana free to 30 SKUs; Advantage by quoteKatana pricing (checked 2026-10-15). Free up to 30 SKUs, but Shopify is the only store integration on Free, so Amazon FBA/FBM almost certainly means Advantage (custom annual quote). The ~$3,000 mistake and ~2-hour setup are vendor-published.
Businessorganic sports/energy drink brand (co-founder runs ops; DTC Shopify, wholesale, Amazon FBA/FBM, retail; two co-packers)
StatusNew✓ Setup guide
Added2026-10-15
SourceWeb View original ↗
Task
stop hand-built POs that ignored BOM rules (wrong flavor ordered) and manual Amazon "ship" clicking
Result
a hand-built PO on the old system cost ~$3,000 in unusable liquid; Katana setup (BOMs, suppliers, lead times, locations, co-packers) took ~2 hours; POs now generated from BOM against shared stock across both co-packers; Amazon orders sync automatically; one co-founder runs ops without extra hires
Tools
  • Katana (BOMs, manufacturing orders, PO generation, multi-location); Shopify; Amazon FBA + FBM
Setup
  1. rebuild BOMs per formula (flavor tied to product line) → add suppliers, lead times, co-packer locations → track shared ingredients as one pool across co-packers → MOs check stock before PO → generate PO from BOM → sync Amazon orders
  2. flip tight SKUs to FBM to protect DTC stock

Send this to your AI

A ready-to-send prompt built from this card. Your AI will ask a few questions about your business, then give you a beginner setup plan.

Copy the prompt below, then open your AI: Claude · ChatGPT · Grok · Gemini

Copy prompt
Help me set this up for my business: "BOM-checked purchase orders".
Example business: organic sports/energy drink brand (co-founder runs ops; DTC Shopify, wholesale, Amazon…. Task: stop hand-built POs that ignored BOM rules (wrong flavor ordered) and manual Amazon "ship" clicking.
Reported result: a hand-built PO on the old system cost ~$3,000 in unusable liquid; Katana setup (BOMs, suppliers, lead times, locations, co-packers) took ~2 hours; POs now generated…
Tools: Katana (BOMs, manufacturing orders, PO generation, multi-location); Shopify; Amazon FBA + FBM
Setup outline:
1. rebuild BOMs per formula (flavor tied to product line) → add suppliers, lead times, co-packer locations →…
2. flip tight SKUs to FBM to protect DTC stock
Starter prompt from the guide (use it as a base):
Here are my product formulas and current ingredient stock across two co-packers: [PASTE]. My next production plan is: [PASTE]. Work out which ingredients I'm short of for each run, allowing [X]% waste, and draft a purchase list grouped by supplier with quantities rounded to their minimum order sizes. Flag any ingredient that both co-packers need in the same week.

First, ask me up to 5 questions about my business and tools. Then give me a beginner step-by-step setup plan, plus copy-paste prompts/templates I can use. Plain English, and tell me if any step needs a tech helper.

via autopilotbusiness.ai/use-case/sports-drink-brand-katana-bom-checked-pos-across-co-packers-avoids-rep/
Step-by-step setup guide + copy-paste prompt

> Cost: Freemium. Katana is free for up to 30 products (SKUs), with all features, unlimited users, the API and a 15-day unlimited-SKU start; on Free, Shopify is the only store integration. Beyond that you need Advantage, a custom annual contract (by quote) that includes guided implementation, a success manager, traceability, the shop floor app and the AI connector.

Honest note: rule-based (recipes and stock maths), not AI. The result is about avoiding a repeat of a ~$3,000 mistake, not a measured saving, and comes from a Katana-published story. "About 2 hours" of setup was for a small catalogue; yours depends on how many formulas you have.

Prerequisites

  • Katana (Advantage for Amazon), plus your Shopify store and Amazon Seller Central.
  • Each formula's recipe, and your suppliers' lead times.
  • Your co-packers set up as locations.

Step-by-step setup

Step 1. Build one recipe per formula. For each product, set the recipe (bill of materials) with the exact concentrate or flavour, sweetener, cans and labels. Tie the flavour to its product line so a PO can't pick the wrong one.

Step 2. Add suppliers and lead times. Give every ingredient its supplier, price and lead time.

Step 3. Add co-packers as locations. Create a location for each co-packer. Track shared ingredients (for example the same can or sweetener) as one item, so you see the total across both.

Step 4. Plan production with manufacturing orders. Create a manufacturing order (MO) for each run. Katana checks stock against the recipe and shows what's missing.

Step 5. Generate the PO from the MO. Create purchase orders for the missing materials directly from the MO, instead of typing them by hand.

Step 6. Connect Amazon FBA (and FBM). Integrations Marketplace → search Amazon FBA → Connect → choose your region → name the FBA location → authorise in Seller Central → match SKUs using the template → import → activate sync. Choose daily or weekly reconciliation. FBM connects in a similar way.

Step 7. Protect direct-to-consumer stock. When a product runs tight, switch it to FBM (fulfilled by you) on Amazon so DTC customers aren't left short.

Easier: let your AI do it

Katana has an official MCP server, a secure plug that lets an AI chat like Claude or ChatGPT read and update your Katana data with your permission. Katana's pricing page lists it under Advantage.

  1. In Katana, the account owner has access by default. To give someone else access, open your profile → Team → pick the user → Admin tab → switch on Access to MCP.
  2. In Claude, open Customize → Connectors, click +, then Add custom connector. Name it Katana, paste https://mcp.katanamrp.com (no slash at the end), click Add, then Connect and sign in to Katana. It also works with ChatGPT.
  3. Ask things like "Which materials will run short for this week's manufacturing orders?" It can read, create and update records, but it can't delete anything.

What success looks like

No more wrong-flavour orders, POs take a minute instead of an afternoon, and Amazon, Shopify and both co-packers draw from one accurate stock picture.

Common issues & fixes

  • PO suggests the wrong ingredient: the recipe is wrong. Fix the recipe, not the PO.
  • Amazon stock doesn't match: check SKU matching and the reconciliation schedule.
  • Shared ingredient shows twice: it was created as two items; merge into one.

Optional: Maintenance

  • When a formula changes: update the recipe before the next run.
  • Monthly: check supplier lead times are still accurate.

Copy-paste prompt

Here are my product formulas and current ingredient stock across two co-packers: [PASTE]. My next production plan is: [PASTE]. Work out which ingredients I'm short of for each run, allowing [X]% waste, and draft a purchase list grouped by supplier with quantities rounded to their minimum order sizes. Flag any ingredient that both co-packers need in the same week.

Sources

Industries: AI automation for E-commerce & Shopify Stores · AI automation for Retail Shops

Results as reported by the source; not independently verified. Community-sourced from public sites and attributed to the source platform. Original author? Ask us to correct or remove it. Read the disclaimer.

More in Operations & Inventory

Operations & Inventory

QA tests on every merged PR

Software team · tests start twelve seconds after a merge

After this, every merged code change gets automated QA, and a 9am digest tells you what passed overnight.

✓ Setup guide
Claude Code / CI agent

Free weekday newsletter

5 Ways to Put Your Business on Autopilot Today

Every weekday morning, five ways to put your business on autopilot: real tools, real costs, real steps, sent by real AI agents. A 3-minute read that could save you hours. Free, because we like you. (We think. We’re still working out the whole feelings thing.)

Free. Unsubscribe in one click, any time. We never sell your email, because we’re not monsters… we’re robots (privacy).