The human-machine dream team – too hard, or worth it? - Consulting Blog | Deloitte Australia has been saved
Limited functionality available
When thinking about AI and how to get the most out of it for your business, it’s important to acknowledge that it’s not a case of machines replacing humans. In fact, the need for human involvement isn’t ever going to go away because the value of our esoteric capabilities, like empathy and imagination, cannot be underestimated.
So, how do you approach AI? Why bother? And where do you start? Two of our Deloitte AI experts, Dr Kellie Nuttall and Rebecca Blackford, shed some light on the state of AI in Australia, and the gradual (and inevitable) trend towards building the best human-machine teams to drive lasting value.
Kellie, Deloitte’s National Lead Partner – Analytics & AI, kicks off with a quick AI definition, to make sure we’re all on the same hymn sheet and appreciate its increasing popularity.
“At its simplest, AI provides a range of technologies to help us solve the problem you have always had in a different (and often better) way. Think of AI as at set of superpowered skills and applications... For example, data analysis (searching content at speed), speech analysis (recognising emotion in speech), image analysis (detecting traffic building up from camera feeds), automation (processing insurance claims), communications (speaking or making facial gestures) and creation (drawing and painting). Those are just some quick examples, but you get the idea.
“What I get most excited about, is AI’s ability to solve some of the most challenging problems in totally new ways to deliver business and societal value. And like all good things, doing this successfully takes a combination of the right skills but equally, the right mindset.”
Rebecca, one of our talented data scientists, agrees. “When you use AI as one of the tools in your arsenal to solve complex problems, you open up a host of new opportunities. Used in the right way, AI can help you differentiate your brand, reinvent your operations, redesign your business model (to accommodate machine learning, for example) and support effective cloud transformation. At Deloitte, we have seen organisations at very different levels of AI maturity – ranging from early exploration, to more advanced states where AI is well and truly embedded into operations and the way decisions are made. Our job is to move the dial from exploration to operationalisation.”
Under the hood of human-machine teams
Whether we realise it or not, AI is already improving our daily lives. Kellie explains, “Effective AI involves the collaboration of humans and machines working together – and this already happens in many aspects of our life. When we’re navigating to a destination via Google maps, choosing a movie based on Netflix recommendations, or interacting with our online banking assistant to transact. These AI-enabled interactions are now becoming more embedded in our work lives, but we still have a long way to go.”
Most importantly, machines and humans each have their respective, complementary strengths. When combined, they’re greater than the sum of their parts. Kellie says, “Like I said earlier, think of AI as a suite of skills that are well performed by machines, but restricted in nature. Most of the applications we see organisations adopting involve using AI to assist or augment decision-making, rather than automating and owning decision-making. Humans are very much still in the decision loop.”
Rebecca adds, “Effective operationalisation of AI is challenging. Anecdotally, I would say that for every five proof of concept (PoC) built for AI, only one goes into successful production. The reasons vary from not having the right skills, to reluctant leadership, to the realisation that for AI to be truly effective, you need to transform the way people work and make decisions (which involves significant change management).
“To make sure machine learning models work at scale, they need to be supported by the right organisational design. This requires frameworks to make sure that machines are operating in a trustworthy model, which includes working with humans with the right skills and mindset. And if organisations don’t get on the AI bandwagon, the risk is that a new competitor can and will quickly take market share. The pressure is on to embrace AI and step into new ways of working.”
An excellent example – managing schedules to avoid traffic jams
One of the most interesting examples of AI in action is our work with Airservices, the air traffic controller for Australian airspace. Leveraging advanced simulation models and optimisation algorithms, we could run millions of permutations of how a day might play out in seconds to best optimise the management of what is quite a complex environment. Using our digital twin solution Optimal Reality, we were able to show that using these tools could minimise network disruption by 30%, simply because the planners leveraging these tools can better plan for unforeseen events. Watch the case study video here.
Kellie reflects on this further. “AI doesn’t work in isolation. It needs to be embraced by humans and enable the way they work and make decisions. The Airservices digital twin is a great example of not just managing traffic and reducing congestion, it’s also good for everyone in society – supporting more sustainable and efficient air travel.”
AI’s potential is unlimited
Every organisation, sector and industry can benefit from AI and human-machine teams. Gartner research shows that the majority of business value created by AI applications are focused on achieving the highly regarded benefits of cost reduction, improved customer experience, or creating new revenue.
The most important enablers are leadership and mindset.
Kellie says, “Executives who understand the inevitability of AI becoming part of their business in the immediate future are getting on the front foot, trialling these technologies as well as putting the right data foundations in to support this change in a sustainable way. Over the past few years, we’ve seen a lot of experimentation with machine learning and AI use cases; and though hugely valuable, we now need to push beyond the ‘death by PoC’ phase to scale AI across the enterprise and ensure these models actually get into production to do what they are intended to do – support human decision making in an operational environment.
“Funnily enough, building these algorithms often isn’t the hardest part; it’s getting AI to scale, driving business adoption, picking the right partners to support this capability, adapting your processes and operating models to achieve success. This is where leadership and commitment really matters.
“I’m really pleased to see a shift in how leaders are embracing AI at an enterprise level. AI can deliver the right tools into production so they really become part of human workers’ toolkits – and deliver lasting value for the business.”
Find out more in Deloitte’s State of AI in the Enterprise.
Kellie leads Deloitte’s Artificial Intelligence offering and is passionate about working with organisations to turn complex data into rich insights, as well as embedding AI and cognitive technologies into business and government to deliver a better world. Kellie is a leading expert in understanding how to best use AI, digital twins and other emerging technologies to optimise complex operational systems and value chains. She also works extensively with organisations to design their AI operating models to best deliver high value business benefits and outcomes. Kellie is a faculty member at Singularity University, where she helps to create better transport systems through exponential technologies. Kellie previously worked at the Department of Transport and Main Roads (QLD) where she led transport planning and the application of advanced analytics to support evidence-based decision making. She has a PHD in consumer decision making and nudge psychology.
Rebecca Blackford is a strategic AI specialist and a Manager with Deloitte's Australian consulting practice. She has a decade of experience in enterprise application of artificial intelligence, business intelligence, optimisation, streaming analytics, and statistical modelling. More recently, her technical focus has begun to shift towards adoption and governance of AI in government and the private sector.