AI shopping assistants: hype vs reality for NZ shoppers

Chat bubbles offering to "find your perfect TV" have popped up on nearly every retail site. Some of that is genuinely useful. A lot of it is a search filter wearing a friendlier costume.

The pitch

Retailer AI assistants promise to replace the tedious work of filtering, comparing, and reading spec sheets. You describe what you want in plain language — "a laptop for uni that's light and has good battery life under $1500" — and the assistant is supposed to narrow the catalogue down to a handful of sensible options, the way a good salesperson would on the shop floor.

That's a genuinely appealing pitch, especially for shoppers who find spec sheets overwhelming. The question is how much of that promise actually holds up once you use these tools for a real purchase.

Where it genuinely helps

For narrowing a large catalogue by soft requirements — "quiet", "good for gaming", "compact" — these assistants tend to perform reasonably well, because they're often just running a natural-language query against the same structured filters (price, category, specs) that a manual filter sidebar uses. You get the same result with less clicking.

They also do a decent job of explaining trade-offs in plain English if you ask follow-up questions, similar to a general AI chatbot. "Why is this one cheaper" or "what am I giving up compared to the pricier model" tends to get a genuinely useful answer.

Where the reality falls short

The gap shows up in three places. First, these assistants are built by the retailer and tend to only search that retailer's own catalogue — so "best laptop under $1500" really means "best laptop under $1500 that we happen to stock," not the best option across the whole NZ market. Second, they can be prone to recommending higher-margin or promoted stock, since the underlying business incentive is a sale, not necessarily the objectively best-value pick. Third, accuracy on genuinely current pricing and stock levels varies a lot between implementations — some pull live inventory, others lag behind.

How to use them without getting steered

Treat a retailer's AI assistant as a fast way to shortlist products within that one store, not as unbiased buying advice. Once you have a shortlist of two or three products, take the model names and check pricing independently across multiple retailers before deciding where to buy.

• Ask it to explain why it recommended something, not just what it recommends
• Note the exact model number, then compare that model elsewhere
• Be sceptical of assistants that push you toward the most expensive option in a category

The bottom line

AI shopping assistants are a genuinely handy way to shortlist products faster than clicking through filters manually. They're not a substitute for comparing prices across the market, since by design they usually only see one retailer's shelf. Use them to speed up the browsing part, then verify the price part somewhere neutral.