‘They barely have wheels, but they do have e-commerce’: What China taught a Kiwi AI expert about AI

Sunday, 22 February 2026

In the middle of having her mind blown on a week’s trip to China recently, Dr Amanda Williamson made a discovery about the country’s use of digital tools that provided something of a salutary lesson for New Zealand businesses.

In the process of walking around Chinese cities, she says she “saw people on the side of the road with barely a cart to hold what they were selling, and yet, they all had QR codes. They had e-commerce, in some cases even without having wheels!”

The average Chinese juice vendor does everything online, as does the consumer, and super-apps like WeChat Pay or Alibaba Pay facilitate not just these payments, but bookings, text and voice messages, video calls, shopping and Ubers. They also feature thousands more apps-within-apps that offer things like ticketing, government services and banking.

And every single street corner and almost any public surface has a power bank to juice up this tech and keep the wheels of millions of daily lives moving.

Williamson was in China to investigate AI and robotics, which is tech that is miles ahead again of what she was seeing on the street corner. But one directly pertains to the other. Because just a few companies in China hold trillions of petabytes and exabytes of data on people through their use of smartphone systems, and from there, they are in prime position to create AI applications that are almost a digital replica of a living, breathing (multilingual) personal assistant.

“For New Zealand, what I found really interesting about all of this is that none of it is novel. We have robots - I’ve seen them in Auckland in the foyer of Sky City. We’ve got facial recognition - I use it every time I unlock my phone. We have sensors, and voice recognition … we have all the components,” Williamson says.

“But what's special is that China is thinking about the data layer and connecting it all up, so you can have this streamlined end-to-end process. There's not a single general manager that I've talked to in New Zealand who hasn't said that their hope and dream is to have a ‘single view’ of customer; actually knowing who their customer is.”

In New Zealand, the idea of large companies owning everyone’s data is not popular, and privacy is one reason systems are still generally siloed - hence the number of stories of recently bereaved widows being offered couples packages and other such corporate faux pas.

There’s a middle way, says Williamson, where you could have the data sorted out safely, and the systems connected up, and then “more easily build out the innovations.”

Fact finding

Williamson is director of the Deloitte New Zealand AI Institute, working with business to explore meaningful ways in which AI can be used in New Zealand enterprise. Her trip to China was done alongside colleagues from around the Deloitte network, and the group explored automation, robotics, AI‑enabled hospitality, and digital commerce infrastructure in Shanghai, Hangzhou, and Shenzhen.

While an expert on moves in the world of AI and how business can benefit from the technology, she is honest about being schooled by the China experience, saying much of what she had been told about the country had been outdated.

“The first thing is that just everything is on a different scale,” she recounts. “I went to cities that I’m embarrassed to say I hadn’t even heard of before … and you realise it’s a city of 32 million people who regularly use self-driving taxis and receive things from delivery drones - not to mention EVs all over, with brands we’ve not heard of.

“As an AI leader, this was wild for me, because I pride myself every day in reading the AI news and knowing what is going on. But we do tend to be oriented to the likes of Silicon Valley.”

China is not quite at the level of US prowess in AI development, but is certainly nipping at its main trade rival’s heels. Just this week, ahead of the 2026 Lunar New Year, Chinese tech giants launched new, open-source AI systems - ByteDance launched Seedance 2.0 video generation, Alibaba introduced Qwen AI 3.5 (which allows models to understand text, images and video simultaneously within one system), and Doubao debuted its 2.0 chatbot. Almost 97% of all Chinese people use AI applications in work and personal life.

FlyZoo

Williamson and her colleagues stayed at Alibaba Group's futuristic 'FlyZoo' hotel in Hangzhou, a 290-room hospitality showcase for its AI innovations. There are no front desk staff in this hotel - guests sign in on a tablet, have their face scanned, and once they’re checked in, the lift scans their face and takes them directly to their room.

Voice commands change the room’s temperature, close its curtains, adjust its lighting (Williamson says she did need to “tag team” with a Mandarin speaker to make it work) and order room service, delivered by a robot. In the hotel bar, a large robotic arm mixes any cocktail you fancy.

Robots are a mainstay of the urban experience in China. Williamson said she started noticing “some very interesting things” including children playing with robot dogs, and shops with a vast array of robots for sale - including exoskeletons for helping people climb mountains, cute robots that sit on your desk and keep you company, and even 'inappropriate' robots sporting huge plastic breasts.

There was also what the Deloitte exec called a “recursive experience” of visiting the robotic mega factory belonging to Swiss engineering firm ABB in Shanghai, China - a 67,000m2 production and research facility where robots produce 10,000 largely bespoke robots every year.

But “walking through the factory, it was very interesting to see what jobs humans were doing,” says Williamson.

“In one place they were coiling cables, and when I asked why, they said humans had the dexterity and precision and careful touch needed for that particular job. In another place, humans were making the frames in which the robots were being sent. And that was because each robot was so variable in size. In another place they were painting the robots, and I asked if they were using (AI-powered) computer vision to do that - and they said no, they were just using the design file.”

Lessons for Kiwis

Williamson came back from China believing there was a thoughtfulness about how the country utilised AI. It was used for certain tasks and not others, with the aim of encouraging business-wide productivity, but humans were still a considered part of the mix.

“They weren't being fooled by how impressive AI is and what it can do; they did not spend money where it was not going to provide a return on investment.”

The Deloitte exec says New Zealand business at present tends to be focusing on “shiny things” like generative AI at a personal productivity level. These tools do not provide the kind of return on investment that those that more fundamentally re-imagine the entire business operation do (tools that, for example, might replace a front desk or project management tasks).

Williamson’s fear for New Zealand is whiplash.

“Yes, we're running so hard at [AI], but …we’re more likely to suddenly realise that, ‘whoa, I spent thousands of dollars on [large quantities of computing power] last year and what did I see for it? What's my measurable impact of this on anything across the enterprise?’

“We need to move it up a level, make it enterprise focused,” she says. And to do that, there’s a level of synching up that needs to happen.

“If we want to move beyond that personal productivity level and actually get enterprise and national productivity gains, which, by the way, we desperately need, then we need to be thinking about getting our data foundations right, safely. Because even if AI didn't progress any more than what it was now, it would still take us years of getting those foundations right before we could get the return which matches the level of innovation that's out there.”