Check your recent online orders for price drops during a sale with Claude

← All use cases / Shopping

ToolClaude

When Amazon's Prime Big Deal Days started, one shopper asked Claude to look through their last 60 days of Amazon orders for anything now significantly cheaper. Five items had dropped, adding up to $83, including $28.59 on baby formula and $20 on a blender, and they shared the prompt so others could do the same.

Prompt to copy
Here are my orders from [SHOP] over the last [NUMBER] days, with the item, date bought and price paid: [PASTE ORDER LIST]. Here are the current prices I've found for the same items: [PASTE CURRENT PRICES OR LINKS]. Make a table showing the item, date bought, price paid, price now and the saving. Sort it by biggest saving first, and only include items that are at least [AMOUNT OR PERCENTAGE] cheaper. Then remind me what I'd need to check in [SHOP]'s returns or price policy before acting on each one.

Replace the [PLACEHOLDERS] with your own details before you send it.

How to try it

  1. Open your order history for the shop and copy the last month or two of orders: the item, the date and what you paid.
  2. Check the current prices, either by pasting the product links or by letting Claude look them up if your version can browse.
  3. Ask Claude for a table of price paid versus price now, sorted by the biggest saving.
  4. For each big drop, check the shop’s returns window and price policy before you do anything.
  5. Decide item by item whether it’s worth returning and rebuying, asking for a price adjustment, or simply leaving it.

Why this works

Big sales often discount the things we bought a few weeks earlier at full price, but nobody has the time to recheck every order by hand. Turning your order history into a simple table makes the savings obvious in minutes. With Black Friday coming, it pairs well with our guide to making a Black Friday shopping plan, so you buy at the right time in the first place.

Tips to get a better result

  • Set a minimum saving, such as £5 or 10%, so you only bother with drops that are worth the effort.
  • Note the last date you can return each item; a saving is no use if the returns window closed last week.
  • Check that the cheaper listing is the same product, size and seller, not a lookalike from a marketplace trader.
  • To keep an eye on spending generally, see our guide to tracking your spending with AI.

Common mistakes

  • Assuming every shop will refund the difference; many won’t, so read the policy before you contact them.
  • Returning and rebuying something you’ve used or opened in a way the returns policy doesn’t allow.
  • Trusting prices the AI hasn’t actually checked. Confirm the current price on the shop’s own page before you act.

Try a variation

Here’s a variation of our own to try next:

Here's my wish list from [SHOP] with today's prices: [PASTE]. Make a table I can update during [SALE NAME], with a column for the sale price and the saving, and tell me which items are usually discounted in big sales and which rarely are.

Related

Read the full guide →

More shopping use cases

Plan an upgrade for an old desktop PC on a budget with Gemini

To upgrade his ten-year-old desktop PC, a Hacker News user listed his current parts, budget and needs for Gemini, which walked him through component options and gave him tips. He spent about £450, spot-checked some of its advice, which held up, and the upgrade was a success.

Source: Hacker News

Find an item in stock at your local shop with a Grok agent

In October 2026 we asked Grok Bot, a desktop AI agent with its own browser, to find a 10 m Cat6 network cable at Anhoch, an electronics shop in Skopje. It searched the shop's website, found an SBOX Cat6 UTP 10 m patch cable for 250 MKD, compared two Lanberg options at 290 and 450 MKD (both warehouse-only) and used the per-branch stock list to find the mall branch that had it. We popped in and bought it in one trip; the agent didn't buy anything itself.

Full guide →

Find a matching wood floor stain from a photo with ChatGPT

Needing to match the stain on wood flooring in a house built in 1999, a Hacker News user uploaded a clear photo taken in bright sunlight. ChatGPT searched online, found a matching stain colour and showed ordering options, and a quart was delivered.

Full guide →
Source: Hacker News

Reference: X @noahkagan