Alan Turing, a British polymath who explored the mathematical possibility of artificial intelligence, suggested that humans use available information and reason to solve problems and make decisions. He pondered the question: why can’t machines do the same? This has become the logical framework of his 1950 paper, Computing Machinery and Intelligence which discussed how to build intelligent machines and how to test their intelligence.
73 years later, our fascination with the possibilities of AI has seen companies spend US$136 billion in 2022 alone. Grand View Research forecasts spending on AI to grow at a compound annual growth rate (CAGR) of 37.3% from 2023 to 2030.
Perhaps a sign of things to come is the US$300 million (dollars) that McDonald's spent to acquire a Tel-Aviv AI start-up, Dynamic Yield, presumably to provide a more personalised customer experience using AI.
Pointing to an IDC report which predicts a 24.5% increase in AI spending in Asia Pacific, C R Srinivasan, chief digital officer with Tata Communications, believes that there is a whole lot of activity that's happening both in the private and in the government space to adopt AI, put it to good use, and to understand how it can be leveraged for various applications, including creating new business models, improving customer experience to name a few.
Critical challenges ahead
For those seeking to find real value in AI, Srinivasan says at this time no AI algorithm can solve the data quality challenge facing enterprises.
Beyond data, other issues include a lack of understanding of AI itself. "There are also talent shortages around AI and insufficient skills. Additionally, skills around being able to adopt open source and integration difficulties with existing systems," he continues.
The connection between the quality of data and data maturity
Srinivasan explains that data quality looks at whether all the data points are captured correctly. He adds that often, there are possibilities of manual errors that can change the quality of data. He cites the example of people keying in incorrect information, which would then affect how AI can process information accurately.
He goes on to comment that data maturity is the ability of a business to quantify the quality of its data. "This includes identifying the processes that contribute to data generation, managing a master database, and ensuring that the data is kept intact," he explains further.
Asked whether it is possible for a business to develop practical AI solutions even if their level of data maturity is not there, Srinivasan comments that it depends on what the organisation is looking for.
"If you’re going to use generative AI to start doing code generation, then data quality wouldn’t be a problem right now," he added. "But if you’re using it for improving things like customer experience (CX), business model, employee retention and productivity, then data quality comes into play."
He cautioned that to ensure that the quality of data used to train the model is accurate and well-validated so that it gives you what you're looking for.
Strategy to help organisations prioritise, build and scale their AI efforts
"It is important to have the right platform – a good understanding of technology, the problems worth solving, and people who can solve those problems," said Srinivasan. He suggested that companies establish a centre for AI with AI-knowledgeable staff or partners.
Also, they must identify key problems and where AI can be used to deliver real value, maximum potential, and scalability. Also, ensure that the technical complexity that AI brings is aligned with existing skills, train people and get business teams involved to scale.
The edge computing paradigm
Srinivasan says edge computing came about because the core cloud systems are a little remote from places where data is being collected. He opines that the more we do work with AI, the more we want to get instantaneous decisions.
"There are many areas where edge computing is relevant and important. Like AI, it is an area that is evolving. I'm sure that it will find good use in the real world," he continues.
Asked to suggest how can CIO's work better with the rest of the organisation to realise how AI can support the business and its goals, and for the IT department to continue to deliver value to the rest of the organisation, Srinivasan opines that CIOs can add value to these conversations by speaking with internal business groups to classify projects and identify key areas of focus.
"This will help them identify the real value that AI can bring to the business, which can then be used to allocate resources and talent effectively," he concluded.
"The ability to prioritise resources to focus on key areas that will derive maximum value, I think therein lies the trick to ensuring that a company has a successful AI program."C R Srinivasan
Click on the PodChat player to listen in greater detail to Srinivasan's perspective on kickstarting the possibilities of AI at the edge today.
- Describe the current state of AI adoption in Asia Pacific and why it's critical for businesses to implement AI.
- For organisations seriously considering deploying AI, what would be one critical challenge that they will have to face and overcome to find real value in AI?
- How does the quality of data relate to the data maturity of an organisation or even an industry?
- Can a business develop practical AI solutions even if their level of data maturity is not there, or will they just be wasting time and resources?
- Can you name one strategy that would really help organisations prioritise, build and scale their AI efforts?
- Could you name one or two business and operational values that edge computing offers as a key advantage for organisations that are considering implementing AI?
- What is the most exciting trend that you are looking forward to materialising in the near term – 2023/24, maybe even 2025/26?
- How can CIOs work better with the rest of the organisation to realise how AI can support the business and its goals, and for the IT department to continue to deliver value to the rest of the organisation?