As of 13 August 2026, AI can only partly optimise your ecommerce product prices.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Export the latest product catalogue, order history, refunds, discounts, stock levels and per-product variable costs from your ecommerce and finance systems, and record whether each price and cost includes VAT.
   2. Open a spreadsheet and make one row per product with product ID, current price, units sold, revenue, refunds, discounts, variable cost, stock level and sales date range, then remove customer names, addresses and other unnecessary personal data.
   3. Add your commercial objective and constraints to the sheet, including the minimum acceptable margin, maximum permitted price change, promotional rules and any products whose prices must not change.
   4. Paste the sheet, its column definitions and the objective into the prompt, then ask the AI to produce the data-quality report and recommendations without filling missing fields.
   5. Check every calculated revenue, cost, margin and percentage change against the source exports, and correct any row where the AI has confused gross and net prices or included refunds incorrectly.
   6. Compare the proposed prices with your live catalogue, current stock, contracts, promotions and current competitor evidence, then ask a colleague to challenge the assumptions and high-risk recommendations.
   7. Apply only a defined test group or a small set of approved changes, record the start date and success measures, and compare the results with the control or comparison period before changing more prices.

   Prompt

   Act as a cautious ecommerce pricing analyst for a UK business. Use only the data I provide and do not invent figures, trends, competitors or customer behaviour. Analyse the attached or pasted product, sales, cost, stock and competitor data and produce:
   
   1. A data-quality report listing missing fields, inconsistent figures, unusual values and assumptions.
   2. Current revenue, units sold, gross profit and gross margin by product, showing the calculation for each.
   3. A price recommendation for each product, with the proposed price in GBP, the expected commercial objective it supports, and a low, central and high scenario where the data supports this.
   4. A clear explanation of the evidence behind each recommendation, separating observed facts from assumptions.
   5. Products that should not have their prices changed because the evidence is weak or the risk is high.
   6. A simple test plan for any proposed changes, including the control group or comparison period, success measures, minimum observation period, and conditions for reversing the change.
   7. A final table with product, current price, recommended price, percentage change, cost basis, expected risk, confidence level and the evidence used.
   
   Do not recommend a price below the supplied total variable cost unless you explicitly label it as loss-making and explain why it might still be considered. Do not treat competitor prices as proof of the right price. Account for VAT only if the supplied data states whether prices and costs include or exclude it. Flag decisions that need a human commercial judgement. End with the five checks I must complete before publishing any price change.
   
   Business objective: [for example, improve gross profit while protecting conversion rate]
   Pricing constraints: [minimum margin, maximum change, promotions, contracts or other constraints]
   Data and definitions:
   [Paste or attach the data here]

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.