The recent release of New Zealand’s ‘wellbeing’ budget received global attention for its focus on more than just the bottom line and economic growth.
While the previous government was focused on fiscal liability in the context of the global financial crisis and two major earthquakes, the new government under the leadership of Jacinda Ardern has shifted the focus to a much broader range of outcomes that lift wellbeing.
To get to this point of enabling a wellbeing approach, New Zealand has been on a journey to better use data and artificial intelligence – or AI – to both understand the challenge as well as provide the tools to coordinate and act on wellbeing.
Benefits realisation with AI
James Mansell has been working with the New Zealand government – from the inside and out – since 2004. He has become a key figure in the government unlocking the potential of data and AI to accelerate coordination, collaboration, manage risk, as well as to respond to demand. This has required engaging with the government and its partners at various levels – from senior ministers to frontline delivery agencies to NGOs.
It has been a journey of discovery for Mansell as well as New Zealand’s public sector.
“When I started off, I was proposing the use of analytics for micro-targeting in very unsubtle ways in the social sector but I soon realised it wasn’t the best way to proceed,” he explained. “It relies on linked data and false positives become very obvious – and it is hard to defend it if done poorly. It’s a high regret option, so I have largely used analytics to work on the performance of the state sector itself – not the people.”
With New Zealand Inland Revenue, Mansell helped to drive the use of AI to deliver value – even after they had invested over NZ$1 billion on replacing their outdated IT system. A program of work was able to demonstrate the value of AI through a series of proof of concepts – including one to illustrate how AI could unlock between NZ$300 to $600 million in additional tax through an AI automated lead campaign. From these results, Inland Revenue is now investing NZ$30 million in its first data lake to automate many routine decisions which will find more tax through smarter campaigns.
AI in government is not always about short term money. Mansell explained it can be used to mobilise a rapid response to an emerging crisis.
After receiving criticism in 2009 on the management of children with harmful sexual behaviours, rapid text mining of child services reports helped the Ministry of Social Development identify more than 400 children exhibiting harmful sexual behaviour within a week – enabling the implementation of services to achieve better outcomes, and reduce the number of their predicted victims by thousands.
But much of James’ work has been to use AI to illustrate the need for a more coordinated and collaborative Government sector and drive reform.
In New Zealand, doctors sign people who are sick or incapacitated onto benefits. And the health sector is the main intervention for getting these people work ready. A 2012 analysis found that approximately 900 doctors accounted for the majority of cost – NZ$3.2 billion worth of benefit payments.
“The challenge here was that the accountability for getting people back to work was sitting with one agency and the ability to achieve it sitting with another,” Mansell said. “This kind of example led to whole of government reform beginning in 2013. Cabinet needed to understand the flow across the systems requiring greater integration between agencies. And that continues today with this government’s recent announcement for its wellbeing budget.”
Pitching AI
In pitching the benefits of AI, Mansell recommends not starting with a “big bang” but smaller projects that can act as a proof of concept for benefits realisation – recognising that there is a lot of levels of maturity agencies may need to go through before they can automate their business processes.
Focusing on the business value is also critical for the AI pitch.
“A lot of people start with the technology, then they find an opportunity, and then they go to business value,” Mansell said. “You actually want to start with the business value, then the opportunity, and then the technology. Most people think of it as a technology challenge, but it’s not – the problem is acting on the insight you have generated. So, if you don’t start with a compelling business value proposition, then you are in trouble from day one.”
And leadership is another consideration, with an understanding needed of how to invest in and act on insights and create agency or system level reform.
“Analytics by its nature is a leadership problem,” Mansell explained. “If you have a genuinely new insight, then you genuinely have to do something different. This is challenging and you have to help people work through that.”
In New Zealand the wellbeing budget is an example of the need for systems reform identified through analytics – a case study that is important for any government agency trying to understand or demonstrate the value of AI.
