This was an opponent that didn’t smile, that wouldn’t engage and that couldn’t keep quiet.
Yet on the long-running popular American television game show, Jeopardy!, it walked away with the first place prize of $1 million.
Represented on stage through an avatar - based off a “smarter planet” logo - this match-winning contestant defeated two of the game’s greatest ever champions, but was nothing more than a few racks of servers.
Yet on this day in 2011 - during a year that saw the passing of Apple founder Steve Jobs, Amazon launch the Kindle and Microsoft buy Skype - IBM Watson first impressed itself onto the wider world, showcasing the beginnings of an emerging cognitive system.
Snowballing into an industry amassing $US8 billion in revenue during 2016, widespread adoption of cognitive systems and artificial intelligence (AI) across a broad range of sectors is turning a one- time gimmick into a viable go-to- market reality.
Yet for partners in Australia, identifying, understanding, and acting on the use cases, technologies, and growth opportunities for cognitive/AI systems will be a differentiating factor.
“Cognitive and AI has been a critical part of business strategy,” Carrington Associates director of technology solutions, Atul Desai, said.
“When we started with IBM we decided to embrace this technology and make this part of our core business.”
Up until 2014, the Sydney-based business intelligence specialist predominantly provided big data analytics solutions to customers across the country, before deepening ties with mining organisations, government and start-ups.
“We realised that the next natural step was to take analytics to mobile devices,” Desai said. “From there, we enhanced our capabilities into machine learning and AI.
“We partnered with IBM three years ago, and started our journey with IBM Watson within the past two years, building and supporting technologies around it.”
As an IBM Watson certified specialist, Carrington is leading the race to cognitive competence, taking huge strides, not minor steps, as the industry gradually matures.
“Software developers and end- user organisations have already begun the process of embedding and deploying cognitive/AI into almost every kind of enterprise application or process, both locally and globally,” IDC research director of Cognitive Systems, David Schubmehl, observed.
Transitioning away from tradition
Recent announcements by several large technology vendors and the booming venture capital market for cognitive and AI start-ups also illustrate the need for organisations to be planning and undertaking strategies that incorporate these wide-ranging technologies.
Spanning government, education, automotive, manufacturing and others, Carrington provides end- to-end projects, training and pre-built solutions, expanding from Oracle focused ERP projects, into business intelligence, mobility, app modernisation and cognitive and cloud through IBM Watson.
Founded in 2006 as a traditional software development partner, Desai acknowledged that despite such progression, the company went through a period of adjustment when learning the difference between how the technology is deployed compared to traditional software.
“Traditional software is heavily reliant on business processes, and about following some standard methodologies to implement a standard or define a new process,” Desai explained.
“[With cognitive computing] the focus is on human interaction, and not the interaction with the system as much but the interaction with the data.”
Ultimately, Desai said value can be derived from making machines do what typically, only humans can do.
“In order to do that, the traditional software approach where you just follow business processes doesn’t work,” he said.
“The key distinction is that these machines continue to learn and provide expert assistance and that is only possible if you keep that engagement going.”
According to Desai - who specialises in business analytics, ERP, enterprise mobility and legacy modernisation practices - the process of training a machine is more “labour intensive” on the part of the provider, but also for the customer which leads to longer deployment schedules.
“It always depends on specific use cases and scenarios but typically the learning can take from one month to sometimes six months,” he said.
Yet Desai stressed that such a process was necessary to ensure the system was more accurate, useful and beneficial to the end-user.
“On subsequent visits, the labour intensity goes down but it is up to the client to ensure the system works for them,” he added.