The Two-Minute Check Before You Press Buy

The basket is full and it's gone eleven. Under the total sits a little box that says "discount code", and you've already decided to buy.

What you want now is a reason not to feel bad about it, and a voice somewhere says there's probably a code.

Plenty of guides promise AI will find every working code and test them for you. Some of that works, now and again.

If you want something that works more often, spend two minutes asking better questions than "is there a code?" Now and then those questions save you money, but more often, I'd say, they save you from buying at all.

Is this a fair price?

Start here, because a 10% code off a price that's been pushed up is worse than no code off a fair price. Turn on web search in the AI, so it can look things up online, and then ask this.

I'm about to buy [exact product, model and size] from [shop] for [price]. Search for what it's selling for elsewhere right now, including the maker's own site. Also check whether it's been cheaper in the last few months. Give me links. Then tell me plainly whether this is a fair price, a good one, or one I should wait on.

Click two of the links before you trust the answer. AI will sometimes quote a price from an old saved copy of a page, or from a different model, and the prices it finds can be hours old. The links are there so you can see the number with your own eyes.

Is there a code that still works?

Lower your hopes first. Coupon sites are full of codes that have run out, and an AI searching them will happily hand you five that stopped working in 2024.

So ask the AI to sort what it finds.

Search for current discount codes for [shop]. Split them into three groups. First, codes you found on the shop's own site or newsletter. Second, codes from coupon sites with a date you can see. Third, codes with no date. Also tell me whether this shop has a first-order discount, a student discount, or a newsletter sign-up offer. Give me a link to where it says so.

The shop's own offers are the ones that work. First-order and newsletter discounts are the most reliable money on the internet, and the most ignored. Try those first, then the dated codes, and don't waste time on the undated ones.

Can't a tool try the codes for me?

Some tools will offer to "test them at checkout", which means a program clicking around a web page with your card details on it. Saving a fiver isn't worth that, so type the codes in yourself.

Do I need it at all?

No code saves more than this question, so ask it before you pay.

I'm about to buy [product] for [price]. Make the honest case against buying it tonight. What do people who bought it say three months later? What's the common complaint? What's the cheaper thing that does 80% of the job? Is a newer version expected soon? Keep it to five lines. Then tell me whether you'd wait.

Sometimes the answer is "buy it, it's fine". Sometimes the answer is "the strap breaks, everyone says so", or "the new one's out in March".

Read "everyone says" as "some people said". Online reviews lean towards the furious and the delighted, and the AI's summary is made from those reviews, so the summary leans the same way.

Either way, you're deciding with the full picture rather than with a countdown timer and a box that says 10% off. The question AI can't answer is whether you'll actually use the thing.

Only you know that, and it's your money.