Trust, not technology, will decide who wins with AI, says Nigel Dalton

Four years into the AI boom, the promised productivity gains had not arrived, keynote speaker Nigel Dalton told the 2026 Proptech Forum in Sydney, organised by Proptech Australia.
Yet, he was optimistic about the future for businesses that use AI to solve real customer problems and earn trust – they are the ones who will thrive in the new AI era.
Thoughtworks social scientist Nigel Dalton said AI adoption had not yet turned into an advantage. He predicted the proptech businesses that would benefit next would use AI to solve real human problems and protect trust.
Delivering his keynote, called The Future with AI, the former REA Group chief information officer said AI use had grown enormously since ChatGPT launched four years ago but the productivity returns had not materialised.
Google now processes around 300 times more AI tokens a month than it did two years ago, he said. “Where's the 300 times productivity lift?” he asked “Where's the 300 times share price? Where's the 300 times happiness? It hasn't happened.”
So far, he said, the gains had gone mainly to a small group of global technology companies.
A trust problem, not a technology problem
Australians were among the world's heaviest users of AI tools, Mr Dalton said. Yet about half say chatbots have made customer service worse, and 70 per cent do not trust businesses to use AI as they see fit.
He said he expected that distrust to bring regulation, including new privacy rules that would require businesses to declare where customer data goes.
"Next year is not going to be the year of profit. It's going to be the year of legislation," he said.
Every new AI tool also adds risk. “What you're really doing is expanding the trust surface of your company,” Mr Dalton said. He pointed to deepfake scams in which finance staff received convincing fake messages from their own chief executive asking them to move money.
What AI can't read
Mr Dalton reminded the audience that people had not changed much in 100,000 years. We avoid change, and we rely on know-how we carry in our heads and pass on through stories. Consequently, AI models could only learn from what had been written down. He estimated that represented only 20 per cent of the knowledge a business runs on.
“We are building an AI-based economy on about 20 per cent of the human knowledge we need,” he said, warning that ideas and data uploaded to large AI platforms could be used to build competing products within months.
“We're not talking about geographic borders anymore. We're talking about intellectual property borders,” he said.
Strong partnerships across an industry were much harder to copy, he added. That made it more important for the leaders of governments and economies to work together.
Ask better questions
Mr Dalton urged businesses not to treat AI as the answer in itself. The lesson so far was that "writing code faster is not the business bottleneck we all thought it was”.
He observed that AI was arriving at the same time as the climate crisis, falling social equity and ageing populations. “Your business opportunity, your customers' problems, all lie in this omelette,” Mr Dalton said.
His advice was to stop asking how AI could transform a whole business. Instead, find one specific customer problem, fix it and then move to the next.
He gave the example of a global software company that added an AI assistant to its product. Customer complaints about the assistant gave the business a ready-made list of improvement priorities straight from its users.
Back to basic human needs
Mr Dalton closed his keynote outlining the needs that have always shaped human behaviour. The property industry and proptech were ultimately in the shelter business, he said. Over the millennia, the need for shelter had never changed. Every society, no matter at what stage of human development, ran on trust. And it was trust that “held them together”.
His keynote set up the Forum's theme, From Adoption to Advantage – Blueprint for Trust in an AI World. The advantage, he argued, would not go to the businesses that adopted AI fastest. It would go to the ones whose customers trusted them to use it well.



