The way we shop is about to change forever – and the consequences are both exciting and terrifying

Tuesday, 17 February 2026

It happened as quietly as a whisper in a boardroom, but the repercussions should not be underestimated for both businesses and New Zealanders looking to hold onto their money and build wealth.

Yesterday, Mastercard successfully completed its first Agent Pay transaction in New Zealand.

Using only AI, Mastercard country manager Megan Simons instructed an AI agent (a bot) to purchase cinema tickets from Event Cinemas on her behalf for Hamnet.

Simons never visited the Event cinemas website, she didn’t enter her Westpac banking app, and she didn’t use Google to find what she was looking for. Every element of the transaction, including date and seat selections, was completed within an AI interface.

In a separate transaction, she booked accommodation with QT Hotels and Resorts in Queenstown for her family.

Using the Matilda interface, she simply asked for a family-friendly hotel in the Queenstown region, within walking distance of the centre, with lake views.

A recommendation popped up, and she hit buy now.

The simplicity of all this is staggering and signals a major shift in how we shop and engage with businesses online.

This innovation is called agentic AI, and it’s typified by having AI complete commercial transactions on our behalf.

Over the last two years, we have steadily been progressing to this point. We’ve trusted AI agents to source information and gather insights on our behalf, removing the need to do Google research independently.

“I’d say probably about 70% of my searches today are probably within an LLM [an AI interface],” says Simons, noting how quickly our behaviour has leaned into the efficiency of AI.

Simons isn’t alone. Estimates suggest nearly 84% of knowledge workers in New Zealand now use generative AI regularly, and a study by Gartner anticipates that traditional search volume will drop by 25% in 2026.

This is all about removing friction and making it easier for people to complete tasks online.

“Not having to trawl through so many websites in order to find what you need is great. It’s become so easy.”

Is it safe?

The widespread application of these features is inevitable, but Mastercard isn’t rushing this technology out the door. It has launched in the US and is only in pilot phase in Australia, the UAE and Latin America.

Simons tells me the plan is to launch it fully in New Zealand within the next six to nine months, at which point any New Zealander could use the product.

The initial demo uses Matilda, but Mastercard is working with all major players, including OpenAI, Microsoft and Google to ensure the experience is consistent no matter which AI interface you use.

Simons tells me that getting the infrastructure right will be integral to ensuring that online shoppers experience a secure environment once the technology becomes widely available.

As things stand, a bot could ostensibly be programmed to do something on your behalf, but the problem is that there would be no transparency.

The retailer on the other side will see a transaction come in and your bank would see the transaction, but neither would know it was orchestrated by a bot.

Mastercard’s plan is to create a system of transparency where both shoppers and retailers are aware whenever an agent is involved in the process. This will help to ensure the legitimacy of transactions and make the ecosystem less vulnerable to fraud.

“We use this phrasing of 'turning the lights on,' so that it's really clear to everybody that this is an agentic transaction,” she says, explaining that transparency will make it easier for a bank to distinguish a legitimate AI purchase from a bot, a scam or a fraudulent actor.

“Trust is the only way to scale this,” she says.

Easier than ever to spend

Victoria Mulligan, the co-founder of Design Futures Aotearoa, advises organisations on the impact technology could have on society in the years to come.

She tells me that this technology makes it far easier than ever before to buy pretty much anything we want.

“I think we’re about to find out what happens when spending money takes absolutely zero effort,” she says.

“Every time we’ve made paying easier, people tend to spend more.

She explains that as we moved from cash to cards and then to contactless and one-click payments, this has systematically removed the “moment of hesitation” that can help to keep spending in check.

With this technology, you could drop in an instruction asking the bot to find you “black jeans, size 30, for under $100 that are well reviewed”, and you could make a purchase instantly without visiting competing websites. So much of the friction that has until now typified the online experience is removed with one simple prompt.

This is a beautifully efficient solution, but it also removes some key decision-making tasks that slow us down and challenge us.

“Making decisions is more than just a practical task,” Mulligan says.

“It's how we stay connected to our own lives. And once we start handing decision-making to AI, we are stepping away from conscious participation in decisions in our lives.”

If we’re not cautious in the way we hand over our decision-making, Mulligan worries that it could affect our ability to discern good information from bad information (or even good jeans from bad jeans).

“It could atrophy over time, and we’re not doing much to help ourselves at the moment,” she says.

You already see this playing out at the university level and among job hunters, where AI slop is often served up as the answer to questions. Some of the information might be accurate, but there are growing concerns that the quality of thought and our problem-solving aren’t being developed in the way that they once were.

If you think this is being over-egged, just consider how many of your friends have become over-reliant on GPS to get to relatively simple destinations.

These concerns are so pronounced that it’s even led to a counter-movement called “friction-maxing”, where individuals purposely introduce friction to combat the frictionless experience of modern convenience.

Examples include flip phones, taking notes by hand, navigating without a GPS or cooking from scratch.

‘Huge’ consequences

Professor Bodo Lang, a professor of marketing analytics at Massey University, peppered the word “huge” throughout our conversation as we spoke about this issue.

“It’s going to be an absolute game-changer,” he tells me.

“We’ll see a seismic shift in consumer behaviour, industry structure and how markets operate.”

He says that once this technology becomes available to the masses, it will take off in no time.

'The two key drivers of consumer uptake with technology are, firstly, the perceived ease of use. So, how easy is the thing to use? And AI is unbelievably easy to use. And the second one is the perceived usefulness. So, does the thing do what I want it to do? And with AI it’s unbeatable in both of those.”

Lang says most shoppers will have no qualms about adopting this tech and will only hesitate if they start to notice glitches during the teething period.

Lang sees the biggest immediate consequences for retailers, who will be vying against each other to become the recommendation that shoppers choose.

Given we have so many small to medium-sized businesses in New Zealand, he anticipates there’ll be a tough period for local firms.

“This will eventually lead to a winner-takes-most outcome,” he says.

“If, for example, I search for that black jeans and it comes back with three recommendations, those then become my consideration set. It is very unlikely that I’m going to discard that and then do my own traditional search.”

So how do New Zealand businesses get into that recommendation zone? And how do they do that better than big firms with far more resources and time to work out how those recommendations work?

These are all big questions that will be hanging over New Zealand retailers as we enter this new era of commerce.

Bodo explains that while the internet gave us infinite choices, AI agents will do the opposite by condensing our options based on the incredibly finite parameters that we’ve set.

Brands that don’t win in the world of AI recommendations could battle to exist.

For entire industries and sectors built on the idea of creating brands that people love, we’re entering a strange phase where a brand might not exist if it doesn’t come up in AI. It’s the modern equivalent of the unseen tree falling over in a forest. Does a brand really exist if no one sees it on AI?

No one is doubting agentic AI is going to be convenient, brilliant and seamless, but we should be concerned about the consequences this creates for both businesses and shoppers as we enter this strange new world where bots do our bidding.

Not to mention the effect this could have on our bank accounts.

What are your thoughts on having a bot do your bidding? Are you excited or worried? Do you think this is a good idea or do you think it will have some serious repercussions? Let us know in the comments. If you’re using the Stuff app on iOS you’ll need to view Stuff.co.nz on a browser to view and post comments.