Written by Isaac Taylor – This article was first published in M2 Magazine
Prepping for a boardroom lunch, veteran event coordinator Anna asked ChatGPT to pair wines with the meals and checked the results with an oenophile colleague. The wines were well paired, but ChatGPT deliberately selected a wine more expensive than the $70 budget Anna gave it. “I was surprised that it is making independent decisions, not just following instructions to combine data,” says Anna.
Anna just learned one of the golden rules of AI, says AI expert, Natalie Rouse, from technology consultancy Mantel Group. “Generative AI models are not always as smart as you expect and don’t reason in the same way as humans. This highlights why humans in the process where AI is making business-critical decisions is essential,” she says.
It is also why Rouse believes that AI will deliver productivity gains without wholesale job losses. While AI has been developed and utilised for decades, the latest wave of AI innovation and exponential capability is sweeping through every sector and reshaping global business landscapes including New Zealand’s. This will be a disruptive time for some, but also means new opportunities and efficiencies.
“All businesses will be impacted in some way or another. Business leaders and boards need to weigh up the risk versus reward of investment in AI, including the risk of doing nothing,” says Rouse. “If we do nothing and our competition creates a 20 percent margin gap through efficiency, or a threefold gain in customer service, where does that leave our business?”
Workers are already finding these gains on their own. Event coordinator Anna uses up to seven different generative AI platforms during her day for tasks ranging from university assignment research, drafting speaker bios and event invites (she always modifies them prior to publishing), and reading out her daily calendar of a morning. The 45-year-old estimates that 10 percent of the written material she produces daily is AI-generated, which she then tweaks and enhances.
Insights from Mantel Group’s State of Data & AI Report suggest that New Zealand is primed to leverage these advancements to take a giant leap forward in productivity. But there might be a lag before we make that leap. The report by Mantel Group, based on a survey of Australasian organisations reveals that while 86 percent of organisations are still in pilot phases or have limited AI adoption, only ten percent have achieved wide-scale adoption.
Historically, New Zealand has lagged approximately two years behind Australia in terms of technology adoption. This time could be different. New Zealand’s tech inertia traditionally stems from the size of the market. “A lot of Australian organisations are dealing with larger data sets and generally larger revenues and budgets. That has allowed them to move earlier in the past,” says Rouse.
Now, things have changed and the rules for tech adoption have been rewritten. “We traditionally thought that data maturity must go in a stepwise linear fashion. You have to get your data or your data platforms sorted out. Then work out your insights. Then you can start looking at predictive modelling. It’s always been that chain you have to follow. But the advent of generative AI has broken that model and levelled the playing field. It is this democratisation that’s got everybody so hyped up about generative AI. Everybody’s got use cases and is thinking how it can be applied in their organisation,” says Rouse.
Productivity gains are just the appetiser
While there are plenty of opportunities across all sectors for AI implementation, the Mantel Group report highlights that 2024 will be a big year, in particular, for the customer service industry. AI technologies, especially those involving predictive analytics and automation, are starting to be leveraged to streamline operations and enhance service delivery, leading to better customer experiences and increased operational efficiency.
“Customer service seems like the first big obvious one because call centres or contact centres are extremely busy usually and can be quite stressful for the operators and can be quite a stressful experience for clients as well,” highlights Rouse.
Organisations are also seeing gains from AI integration in strategy, finance, product development, and marketing and sales. Of particular note is the value organisations are finding in employing data and AI to predict consumer behaviour. A standout example of practical AI application is detailed in Mantel Group’s collaboration with Woolworths NZ. Woolworths NZ’s data science teams were initially hampered by siloed operations and a lack of standardised processes, which stifled their ability to innovate rapidly.
Mantel Group addressed these challenges by introducing a standardised machine learning (ML) project template and delivery practices that could be applied across any advanced analytics scenario. This strategic overhaul led to significant improvements for Woolworths NZ, including accelerated development cycles, reduced dependency on manual maintenance, enhanced data and model quality, and overall better outcomes from data science initiatives.
“Woolworths grasped two of the most fundamental principles of successful AI adoption. The first is that AI hinges on good quality data. This is one of the single greatest challenges organisations face – understanding what data exists, who owns it, its relevance and quality,” says Rouse. “The second is that they considered scale early. It’s never too early to think about scale, but it can be too late. Running controlled AI experiments to learn about what works and what doesn’t, and to validate that your controls work, is the right approach. But if you don’t build with scale in mind from the get-go you might quickly run into barriers.”
Freeing up humans to do more human things
AI’s ability to handle routine inquiries, speed up response times, and free up human agents to tackle more complex issues has ramifications across the board, suggests Rouse. “It’s quite interesting that all the big headlines are about how generative AI can replace the creative side of the human experience like generating language, art, poetry, and music. We saw this with the first wave of automation, people thought if 30 percent of a certain job was automated they would directly reduce the workforce by 30 percent, but it just never paid off in that way.”
“Really, where the rubber is hitting the road in terms of increasing productivity is automating away jobs that humans don’t like. They might be repetitive. They’re probably relatively easily well-defined. For example, translating highly technical language into a more human-readable language. That sort of thing is pretty straightforward for a large language model and pretty mind-numbing for a human. Anything that’s a high cognitive load in terms of quantity rather than quality can be really well-solved by a large language model. That frees the human up to do the more creative, more enjoyable tasks. That’s the real productivity gain we’re seeing,” says Rouse.
New business models are the main course
Productivity gains, however, are only half the story. The real value of AI will be unlocked when companies take advantage of the capability as a competitive advantage, using AI to improve existing products and create entirely new products and experiences for customers.
This next wave of AI innovation will thrust IT departments further centre stage, suggests Rouse, making them central to business innovation within organisations. However, this hinges on the relevance of AI investments to the fundamental challenges and objectives of the business. Strategic alignment of AI projects is currently lacking, Mantel Group’s report shows.
“It all stems from the overarching business goals. How are we making our organisation better? What are our goals? What are we trying to drive? Is it customer experience? Is it revenue? Is it all of those things? What are the things that we need to do to move the needle on? How does technology support that?” says Rouse.
Balancing risk Vs reward
Rouse also points out the need for robust data and AI governance in the integration and operational phases of technological advancements within businesses. She emphasises that while generative AI, like large language models, presents a breakthrough in how easily it can be applied to business processes, the ease of implementation also brings heightened responsibility.
Rouse explains that generative AI has “jumped the fence” from being a complex technological concept to a practical tool readily accessible for business applications. This accessibility means that businesses don’t need a deep understanding of the underlying models to start applying AI to enhance operations, such as automating content creation. However, this ease of use should not bypass the need for security considerations and governance.
Governance and security might be seen as a kind of innovation limitation, but in reality, they are crucial for safe and effective AI implementation, highlights Rouse. Proper governance ensures that businesses are aware of and can manage the risks associated with AI technologies. This includes having clean, well-managed, and secure data, as well as a thorough understanding of how AI applications could impact the business and its stakeholders. Good governance practices should enable innovation by providing a framework within which risks are managed.
“I think some people see governance as kind of a dirty word sometimes, but it absolutely isn’t. If you don’t have those right governance processes in place, you can really open yourself up to risk very easily. There’s data governance and then there’s AI governance, and AI governance builds on and extends data governance. You need more than just good data governance to ensure that you’re understanding the risks properly, that you’re going through the right processes. Good governance should be about enablement, not about restriction.”
Enablement is what we need as New Zealand businesses explore the capabilities of AI. But as Rouse suggests, exploration and adoption require a balanced approach that includes robust governance and security as well as a clear alignment with business goals. If we get it right, AI can be a powerful catalyst for growth.
