1. ChatGPT Won't Source For You. But It Will Think With You.

Most sellers use ChatGPT wrong. They type "find me profitable Amazon products" and get a list of generic nonsense they could never source. That's not how this works.

ChatGPT is a thinking partner. It helps you research faster, build systems, check your logic, and prep your VAs. Used that way, it saves real hours every week at zero extra cost.

I run a $100K+/month Amazon FBA operation and I use ChatGPT daily. The prompts in this post are ones I run, not theoretical examples. If you want the full picture of how AI fits into the larger scaling system, start with how to scale Amazon FBA.

These 15 prompts are grouped by use case: lead generation, product vetting, margin math, SOP building, brand risk, and VA training. Copy them, paste them, and modify the bracketed placeholders with your real numbers.

2. The One Rule That Makes Every Prompt Work

Garbage in, garbage out. The biggest mistake sellers make is prompting with no context. "Is this product good?" gets you nothing useful.

"My buy cost is $9.40, Amazon sells for $24.99, FBA fees are $6.20, there are 8 sellers, BSR is 12,000 in Home & Kitchen, should I buy 30 units?" gets you a usable analysis.

Add numbers. Add constraints. Tell ChatGPT what you already know and what specific decision you're trying to make. The more context you give, the more specific the output.

"Doing 227 billion in revenue with 4,000 SKUs... that's around $50 million per SKU. And I know that there's a lot of Amazon sellers that basically have the same amount of ASENs in all storefronts... it is not the right way to run a business." Chris Mangunza, AMAZON SELLERS: LOOK AT WHAT COSTCO IS DOING (Nov 2, 2023)

Costco made $227B on 4,000 SKUs by staying focused on a tight catalog. The same principle holds for prompts: give ChatGPT more context and you get a more useful answer.

3. Prompts 1-3: Generating Product and Category Ideas

These prompts help you decide where to look. ChatGPT doesn't know today's retail prices or current Amazon BSRs. Use it to identify the right stores, departments, and categories before you open a sourcing tab.

Prompt 1: Store department targeting

"Act as an Amazon FBA sourcer. List 15 store departments at Walmart and Target where clearance items most often have high Amazon demand. For each department, explain why it works and what seasonal patterns I should know about."

Prompt 2: Sub-category brainstorm from your niche

"I sell household and kitchen products on Amazon FBA using online arbitrage. Give me 20 related sub-categories I might be missing that have strong repurchase rates, low IP risk, and products that typically retail under $30. Explain the rationale for each."

Prompt 3: Lateral product discovery

"I found this product: [product name]. What competing brands, bundle variations, or related SKUs should I check on Amazon to find similar sourcing opportunities at similar price points? List at least 10 directions I could research."

These prompts prime your research so you're not staring at a blank screen. For faster lead generation, check how to build an OA buy list fast and browse the best OA lead lists in 2026.

4. Prompts 4-5: Pre-Buy Product Vetting

Before you place an order, stress-test the deal. ChatGPT pokes holes in your reasoning if you give it enough context. Ask it to find problems with the deal. That's the most useful thing it does at this stage.

Prompt 4: Full deal analysis

"Analyze this OA deal for me. Product: [name]. Buy cost: $[X]. Amazon sell price: $[Y]. FBA fees: $[Z]. Number of FBA sellers: [N]. BSR: [rank] in [category]. Price history: [stable/volatile/declining]. Tell me the risks I'm missing and whether this is worth buying 30 units of."

Prompt 5: Sales velocity estimate

"The Amazon BSR for this product is [rank] in [category]. There are [N] FBA sellers competing on the listing. Based on typical category sales velocity, estimate how many units per month I could realistically expect to sell as one of those sellers. Walk me through the math and state your assumptions clearly."

Prompt 5 produces estimates. Think through velocity before you commit capital. Pair the output with real Keepa data for every serious buy. If you're new to Keepa, here's how to read Keepa graphs.

5. Prompts 6-8: Running the Numbers

Run your margin math in SellerAmp or the Amazon FBA Calculator. Use ChatGPT to build reusable frameworks and check logic fast. Keep the verified tools for final numbers.

Prompt 6: Quick profit check

"My buy cost is $8.50. The current Amazon selling price is $22.99. FBA fulfillment fee is $5.80. Referral fee is 15%. Inbound shipping is $0.45 per unit. Prep and labeling cost is $0.50 per unit. Calculate my net profit, ROI, and the minimum selling price I need to hit 30% ROI."

Prompt 7: Build a reusable margin template

"Create a margin calculator template I can paste into ChatGPT for every product I evaluate. Inputs: buy price, sell price, referral fee %, FBA fulfillment fee, inbound shipping per unit, prep cost per unit. Output: net profit, ROI percentage, profit margin percentage, and minimum sell price to hit exactly 30% ROI."

Prompt 8: Q4 ROI logic check

"Explain what minimum ROI I should accept on fast-moving vs. slow-moving products during Q4 (October through December) compared to Q1 and Q2. Walk me through the reasoning around holding costs, storage surcharges after 181 days, and velocity differences by season."

For more on this, the FBA profit margin breakdown goes deeper on what numbers matter when you're scaling past $10K/month.

6. Prompts 9-10: Building SOPs You'll Actually Use

One of the highest-value uses of ChatGPT is drafting standard operating procedures. You describe your process, it writes the steps. You edit, approve, and hand it to a VA. First draft in 90 seconds instead of 90 minutes.

Prompt 9: Keepa SOP for VAs

"Write a step-by-step SOP for evaluating an Amazon product using Keepa. Include what to look for in price history (at least 90 days), BSR history, and number of offers over time. Flag the signals that indicate a product is risky. Format it as a numbered checklist a beginner VA could follow without additional guidance."

Prompt 10: Pre-buy vetting checklist

"Create a 10-question vetting checklist I can run before placing any online arbitrage order. The checklist should help me avoid IP issues, rank instability, coupon clipper traps, and brands that restrict resellers. Keep it short enough to complete in under 2 minutes per product."

ChatGPT drafts these fast. You refine them with your own sourcing experience, add brand-specific rules, and publish them inside your operation. See how to build FBA SOPs for a full framework on documenting your business.

7. Prompts 11-12: Brand Vetting and IP Risk

Brand risk is real and it's getting worse. Nike sent copyright infringement complaints to resellers in late 2025. Listings dropped from 45 to 23 sellers overnight on specific ASINs.

"Nike is sending IP complaints. A ton of sellers have received IP complaints from Nike. Not IP, but copyright infringement complaints, which are pretty much the same thing... this listing went from 45 to 23 sellers." Chris Mangunza, Amazon Sellers PANIC as Nike Sends Copyright Complaints (Nov 11, 2025)

Brand vetting before you buy is non-negotiable. ChatGPT can help you build a quick screening process that runs in under 5 minutes per brand.

Prompt 11: Brand risk research checklist

"I'm thinking about sourcing products from [brand name] on Amazon. What specific searches and checks should I run to determine if this brand actively sends IP complaints, enforces MAP pricing, or has restricted third-party resellers in the past? List the exact steps in order."

Prompt 12: Universal brand vetting prompt

"Act as an Amazon IP risk analyst. I'm about to source products from [brand name]. Based on what you know about this brand's history with Amazon resellers, give me a risk rating of low, medium, or high. List the main concerns and tell me what I should verify independently before buying."

For the full list of brands to stay away from, check which Amazon brands to avoid in OA.

Free Live Training

Watch me run this system live every Thursday

Every Thursday at 8 PM EST I run a free 60-minute training where I source, analyze, and ship a real product. Reserve a seat and watch the whole thing.

Reserve My Free Seat →

8. Prompts 13-15: Training VAs with AI

Once your sourcing process is solid, get someone else to run it. ChatGPT helps you document your criteria in plain language so a VA can apply them consistently from day one.

Prompt 13: VA sourcing training instructions

"Write training instructions for a virtual assistant on how to identify online arbitrage leads from Walmart.com. Include: what metrics to record for each product, what screenshots to take, how to submit leads for review, and what criteria automatically disqualify a product before it even reaches me."

Prompt 14: VA deal feedback template

"My VA submitted this product for review: [product name, buy cost, sell price, BSR, number of sellers, Keepa notes]. Write a feedback message explaining whether it passes or fails my sourcing criteria and exactly why. Keep it educational so they improve from each rejection."

Prompt 15: VA interview questions

"Write 10 interview questions to screen a virtual assistant for Amazon online arbitrage sourcing. Include questions that test whether they understand basic margin math, can identify red flags in Keepa graphs, know what FNSKU labeling means, and have sourced from US retail sites before."

These prompts get you a first draft fast. The real work is iterating with your VA over the first 30 days. See how to train a VA to source for you for the full onboarding process, and compare the tradeoffs in AI vs. virtual assistant for OA.

9. What ChatGPT Gets Wrong

ChatGPT doesn't know today's prices. It doesn't know current Amazon fees for a specific ASIN, real-time BSR for a live product, or whether a brand started gating categories this week. Never use it as your primary research tool for live data.

It also makes math errors. Verify any numbers from Prompt 6 and Prompt 7 against SellerAmp or the Amazon FBA Calculator before you place an order. ChatGPT confidently produces wrong numbers sometimes.

The biggest risk is confirmation bias. If you tell ChatGPT "this product looks great, right?" it will usually agree. Ask it to find the flaws. Phrase it as: "Tell me every reason this deal could go wrong."

Use it for thinking, drafting, and building frameworks. Run final numbers through SellerAmp or the Amazon FBA Calculator.

10. Pairing These Prompts with Real AI Tools

These 15 prompts run in standard ChatGPT free or paid. They work right now with no setup. Sellers scaling past $50K/month use AI tools that connect directly to Amazon Seller Central data, Keepa graphs, and live pricing.

To see what that level looks like, check out the best AI tools for Amazon FBA in 2026. For a direct comparison of AI assistants, Claude vs. ChatGPT for Amazon sellers breaks down which one handles which tasks better.

The prompts in this post are your starting point. You'll refine them as you use them. Every time a prompt gives you a bad output, ask yourself what context you left out and add it next time. Within a few weeks you'll have a custom prompt library that fits your exact sourcing criteria.

Sellers inside The Scaling Society get access to our full prompt library with over 60 tested prompts, plus the systems to run them at scale. To see how the whole operation works, grab a free seat at Thursday's training and watch me source a live product from start to finish.

11. Next Steps

Prompts cut research time. The posts below build the system around them. Read in order: