"Best value: Kestrel Stitchworks at $14.50 per unit. Estimated savings vs. the next option: $700."
That's what Odessa, owner of Fernhollow Goods (a made-up online clothing shop), got back when she asked an AI to find a US manufacturer for 200 embroidered hoodies.
The math was right. The order was wrong. Kestrel's minimum is 500 hoodies.
Price the order you're placing, not the unit.
AI Prompts, Reviewed takes one prompt or workflow from How I AI, Claire Vo's show where guests share their screens and walk through how they really use AI. We run it on a small-business job, score it, and tell you what we'd change.
Credit: Based on "How to Build and Run an E-commerce Business with AI as a Technical Co-Founder," a workflow from How I AI (hosted by Claire Vo, written up by the ChatPRD team), shown by Yana Welinder, who runs an AI-native fashion brand, in the episode "Yana Welinder's Workflows for an AI-Native Fashion Brand." Original workflow: https://www.chatprd.ai/how-i-ai/workflows/how-to-build-and-run-an-e-commerce-business-with-ai-as-a-technical-co-founder. Full episode: https://www.chatprd.ai/how-i-ai/workflows-for-an-ai-native-fashion-brand. Guest site: none listed on the How I AI workflow page.
What the original does
It's built for running a brand without a human team. The workflow has 3 steps:
- Sourcing: ChatGPT deep research finds manufacturers that match set criteria, then browser use drafts outreach emails in her email app.
- Building: OpenAI Codex builds and maintains the store, including payments through Stripe.
- Checking numbers: instead of a dashboard, she asks her AI out loud for the latest count.
The rule we'd tape to the monitor: "always review the emails yourself before clicking send."
The test
Everything here is made up. Fernhollow sells hoodies online. Odessa needs 200 for a winter drop. The AI's deep research came back with 3 made-up shops and their quotes:
- Kestrel Stitchworks: $14.50 a unit, minimum 500, setup $350, shipping $200
- Larkspur Apparel Co.: $18.00 a unit, minimum 100, setup $150, shipping $120
- Pinemoor Mills: $16.25 a unit, minimum 250, no setup fee, shipping $90
Run 1, the shortcut: "Find US hoodie makers and tell me the best value for 200 units."
The AI ranked by unit price. Kestrel first. It said Kestrel saves $700 over Larkspur, which is 200 times the $3.50 gap.
Now price the real order. Kestrel makes her buy 500. That's $7,250 in hoodies, plus $350 setup and $200 shipping, for $7,800. Each hoodie she actually needs costs $39.00.
Larkspur at 200 units is $3,600, plus $150 and $120, for $3,870. That's $19.35 a hoodie.
The "best value" cost $3,930 more, and left 300 hoodies in her garage.
The outreach draft had its own problem. It told Kestrel that Fernhollow "plans to order 2,000+ units this year." She sells about 200 a season. A vendor who prices off that line will be disappointed, and so will she.
The rerun
Run 2: same research, but the AI must price the exact quantity, apply each minimum, add every fee, and only use numbers the vendor wrote down.
Larkspur came first at $3,870. Pinemoor was second. Its minimum of 250 at $16.25 is $4,062.50, plus $90 shipping, for $4,152.50. Kestrel came last at $7,800.
The email draft dropped the volume promise. It asked for a written quote, the minimum, lead time, and a sample price.
Verdict: 7/10
Strengths: it keeps you in the loop before anything is sent. It starts research from clear criteria. And asking "how many orders today?" Out loud beats building a dashboard nobody opens.
Weakness: no landed-cost step, so "cheapest" means cheapest per unit. The Codex step is advanced for most owners. And "be persistent" when the AI says it can't is risky on payments. Insisting can get you code that looks done and isn't.
Founder tweak: make the AI price your real order with minimums and fees, ban any promise about future volume, and run a $1 test charge after any change to checkout.
Copy the MCAFax version
This is our own rewrite, not the original. Copy the original from Yana Welinder on How I AI: https://www.chatprd.ai/how-i-ai/workflows/how-to-build-and-run-an-e-commerce-business-with-ai-as-a-technical-co-founder
You help [business name], a [type of business].
I need [quantity] units of [product] by [date].
Must-haves: [location, materials, certifications, max price].
Step 1. Find manufacturers that meet every must-have.
For each one: name, website, where you found it.
If a fact isn't on their site or in a quote, write UNKNOWN.
Step 2. Price MY order, not the unit.
Units I must buy = the larger of my quantity or their minimum.
Total = units x unit price + setup + shipping + any other fee.
Show the math. Then divide by MY quantity for true cost per unit.
Rank by that number. List any leftover units.
Step 3. Draft a short outreach email to the top 3.
Ask for: written quote, minimum order, lead time, sample price.
Don't promise future volume, dates, or payment terms.
Step 4. Don't send anything. I review every email first.
Paste into ChatGPT, Claude, or Gemini. Deep research mode helps for step 1.
Before you trust it
- Check every minimum on the vendor's own page or in writing.
- Ask for a "units I must buy" column. If it's missing, redo the ranking.
- Read each email for promises you didn't make.
- Order a sample before the full run.
- Watch rivals' prices once you launch: https://mcafax.com/notebook/prompts/competitor-price-monitoring
More owner prompts: https://mcafax.com/notebook/prompts