You are here

OPINION

Define goals before taking the digital plunge

BUSINESSES in Singapore are discovering new applications of artificial intelligence and machine learning (AI/ML) even as concerns remain over the ability of an organisation to adapt to these digital technologies and be future-ready.

AI/ML technology capabilities such as image learning, speech recognition or text analysis are leading companies to discover not just solutions to existing challenges but also unearth new problems they did not know had existed and explore new business opportunities.

AI has helped businesses drastically cut costs, increase revenue and optimise business operations. The successes that the early adopters of AI/ML technology have experienced, coupled with the increased sophistication of these technologies, have inspired Singapore businesses to take the plunge. However, for end results to match expectations, business leaders should clear the fog that often surrounds AI and develop a clear understanding of its capabilities.

A decision to deploy AI/ML must start with answers to these questions:

sentifi.com

Market voices on:

  • What specific business problems must we address?
  • Will the business benefits justify the costs?
  • Do I have sufficient data to feed the AI engines and train the networks?
  • Do I build my own in-house team or partner with a solution provider?

Uncovering limitless possibilities

We have seen that once a company has identified the business problems it wants addressed, it needs help in defining the strategy and designing and building a solution. For instance, a common problem for a retailer is fraud at a self-checkout point: a customer scans a low-value item but places a high value item in the bag. A retailer may want to deploy deep learning via image recognition to trace a customer's shopping activities and catch fraudulent behaviours.

But with the right strategy, the company can also identify other potential areas of revenue generation. It can build a neural network with the help of data from wearable devices, sensors, location sensing wi-fi and mobile apps. And then add a layer of deep ML learning to capture all the micro moments in the customer's shopping journey. The benefits would be far greater: deeper customer engagement, better conversion and higher revenue.

We collaborated with Singapore-based developer of AI solutions, Xjera Labs, on a proof-of-concept for a video analytics solution to make the workplace smarter and safer. The solution, which has been developed for a food processing plant, analyses data from cameras at the plant's entrance to ensure workers follow hygiene and safety rules. If the camera detects a worker without a hair net or helmet, it sends an automatic alert to the authorities. With this solution, the company can do away with additional personnel and improve the safety and hygiene levels in the plant.

The future-readiness of Singapore businesses

Earlier this year we released two reports - the Fujitsu Future Insights Global Digital Transformation Survey Report 2018 on the current state of AI initiatives and implementation challenges across industries and countries, and the Digital Transformation PACT (People, Actions, Collaboration, Technology) report on the digital readiness of organisations.

The first report, based on a survey conducted among 1,535 C-level executives and decision-makers in 16 countries, shows that business leaders understand that AI is set to have a significant impact on their industries, organisations and wider society. A majority of the respondents are positive about its future development, with 68 per cent believing that the future will involve people and AI working collaboratively.

The Digital Transformation PACT report, based on a survey of 1,625 global business leaders, found that globally one in three (33 per cent) has cancelled a digital project in the past two years at a cost of close to S$ 700,000, while one in four (28 per cent) has experienced a failed project costing over S$ 900,000.

Among all the survey respondents, Singapore business leaders are the most confident about implementing digital transformation initiatives including AI and ML. Fewer than half of Singapore businesses said that the fear of failure seriously hindered their organisations' digital transformation initiatives, while nearly 2 in 3 business leaders said that digital is creating entirely new business processes and functions - the highest among all countries surveyed.

However, several concerns remain among Singapore business leaders about future-readiness. Over two-thirds (70 per cent) are concerned about their organisations' ability to adapt to technologies such as AI. Nearly all (98 per cent) believe upskilling of their existing staff will be vital for success in the next three years and 94 per cent feel by 2020 AI will transform the skills needed internally.

AI development: what should you expect?

Our experiences in the field of AI research through our own Riken AI Deep Learning Environment (RAIDEN) have shown us the importance of building enormous neural networks for deep machine learning. These networks are being built with massive increases in computational capacity, data volumes and algorithm complexity. As neural networks grow, we will see AI producing more accurate results and handling more complex tasks.

High performance computing will be critical for AI development. It works by taking large, complex tasks and splitting them up into many smaller tasks that are solved separately. There is technology emerging in this space that will make it possible to apply this approach to neural networks.

To accelerate AI and build faster neural networks, large memory GPU systems are required. The memory hosts a larger neural network, and the faster GPU will accelerate the vector computations required for neural network processing.

And lastly, most of the recent AI development are centred around data-driven algorithms, such as multiple-layered convoluted neural networks. Massive amount of data is fed to the software to learn patterns and behaviour. This means the data is more important than the algorithms and hence easy access to algorithms and computing power is the key to drive adoption. This has led to many technology vendors pushing AI in open source software development, such as Tensor Flow (from Google), Cognitive Toolkit (CTK from Microsoft), SageMaker (from AWS), and many cloud providers are enticing users to develop on their cloud platforms using their tools for automation. Hence open source software and cloud computing will become the logical platform for AI development.

Digital is the future of business and AI will be a defining technology in that mix. There is no better time than now for organisations to make the transition. But before making any investment, business leaders must define the business goals, understand the capabilities and compatibility of the various parts that make the digital core, and form the right partnerships for a successful digital transformation.

  • The writer is country president of Fujitsu Singapore.