Can AI actually tell you the best time to buy? We tested it

We asked several AI tools when to buy a handful of common electronics categories, then checked the advice against actual NZ price history. The results were a mixed bag.

The test setup

We asked general-purpose AI chatbots and a couple of dedicated price-tracking tools with "AI" branding a simple question for several product categories: "should I buy now or wait?" We then compared that advice against the actual, observable price history for equivalent products over recent months, plus what we know about NZ's recurring sale calendar (Black Friday, Boxing Day, EOFY, and periodic clearance events).

This wasn't a rigorous scientific study — more a sense-check on how much weight this kind of advice deserves for an everyday shopper.

Where the advice tracked reality

For big, predictable categories — TVs, headphones, established phone models — the general shape of the advice was often reasonable. Tools correctly flagged that prices on older TV models tend to soften in the months before a new model year lands, and most correctly pointed toward the well-known NZ sale windows as good times to watch. This kind of advice is really just applying well-known seasonal patterns, which doesn't require much sophistication to get right.

Where it fell down

The advice got noticeably shakier for newer or niche products with limited price history, and for anything where the "wait" recommendation depended on an unannounced future event. A few tools confidently suggested waiting for a specific upcoming sale event without being able to verify that event was actually confirmed — sometimes conflating rumoured dates with confirmed ones. General chatbots, in particular, occasionally cited discount percentages or price points that we couldn't verify against any real retailer data, which is a classic sign of the model generating a plausible-sounding but unverified answer.

What actually worked best

The most reliable signal wasn't any single AI prediction — it was looking at the actual price history chart for the specific product and cross-referencing it with the NZ retail calendar ourselves. Tools that showed their historical data alongside the recommendation (rather than just a verdict) were much easier to sanity-check, and we trusted those more.

• Seasonal advice ("wait for EOFY") tends to be reliable
• Product-specific percentage or dollar predictions deserve more scepticism
• Always check whether the tool shows its underlying price history, not just a verdict

The bottom line

AI-driven buy-timing advice is decent at the broad strokes — seasonal sale patterns most shoppers could learn themselves — and weaker on specific, confident predictions about individual products. Use it as a nudge toward the right season to buy, then verify with an actual price history chart before pulling the trigger.