The Path to Enterprise AI: IDC Reports Variety in the APAC
Few technologies have quite managed to fire the imagination of business decision-leaders in recent years as much as the possibilities of artificial intelligence have.
In the white heat of the APAC competitive marketplace, AI has the capability to completely change the game for companies that are quick to adopt. In many verticals, a new wave of innovators and disruptors could come from anywhere. The only thing that’s 100 percent sure is that new generation of game-changers will be using AI right across the organization, in every business function.
However, an IDC InfoBrief, sponsored by Dataiku, shows that in the Asia Pacific region, there are very different rates of progression in AI adoption, with variations in geography, industry type, and distance down the path traveled towards full AI adoption.
The aims and objectives of using AI technology with regards to business practices are numerous, but there are several trends emerging, irrespective of geography. Business executives right across the region are agreed that machine learning and AI modeling can be used in all areas of an organization, but especially for driving customer engagement, increasing productivity, and spurring product innovation.
But before we take a deeper dive into the IDC InfoBrief’s findings, it’s worth noting that in 2017, 60 percent of executive decision-makers stated that they believed AI would be a crucial requirement for their businesses to gain a competitive edge. But two years later, in 2019, the same organizations reported that only 26 percent of them had a consistent enterprise-wide AI strategy, and a mere seven percent had AI projects that consistently met expectations.
In this article, we aim to look at some of the reasons why companies in APAC are struggling to deploy enterprise AI in ways that are making a significant difference, and also point to some of the findings as to how this situation can be rectified — albeit, the majority of the answers are better found in, and digested by reading the full document.
The trio of requirements
Any enterprise adopting AI as an integral part of its overall business strategy requires three elements: data, people and technology. Firstly, data has to be acquired, normalized, and managed throughout its lifecycle to support quick decision-making. At the same time, all information has to keep to local and international governance with regards to security, transparency in practice, and privacy.
In businesses in the Asia Pacific, on average only 20 percent of businesses are leveraging data acquisition- and preparation-as-a-service. That means the real-time benefits go wanting because the necessary tools are not in place. One sector that ranks much higher on this particular metric is the region’s manufacturers: that may be due to the long-establishment of interconnected IoT devices in industry, where operational technology has been effectively bedding-in for over 30 years.
While the paper shows that many enterprises have assembled AI-skilled staff into COEs (centers of excellence) — often led by a chief AI officer — 51 percent of AI personnel are distributed and uncoordinated across the various business functions.
What the InfoBrief also shows is that the AI platforms are best understood as collaborative hubs. Such hubs involve business decision-makers, IT staff, data scientists and stakeholders with deep, local knowledge. Among the market leaders, that joined-up approach is widely acknowledged as producing game-changing results.
The third prerequisite for effective AI in enterprise is, naturally enough, technology. At all levels of the journey to Enterprise AI (from experiment, through multiple deployments, to constant and proactive disruption powered by AI), the choice of technology platform is a critical enabler.
The study shows that 77 percent of AI models remain in the pre-production phase in the region, with even the most successful countries’ organizations (in South Korea and China) only having AI models in constant production in 30 percent of cases. Hong Kong has the lowest uptake of technology solutions specifically designed for AI with only 15 percent of businesses having AI effectively in production as an integral part of the business.
Data the least problematic
Of the three areas touched on, above, it’s data — acquisition, normalization, modeling, and source of insight — that is least problematic, the study shows. Data-related issues only become more important with increased scale. (Even then, the constant disruptors’ use of AI, the most advanced organizations, only experiences issues in data collation and processing 19 percent of the time.)
The full story
AI can help enterprises disrupt their business and become an important part of their strategy, but it takes time and commitment. They begin, naturally enough, with looking at best practice. That means developing an AI COE (center of excellence), enhancing their data management workflows, and creating an AI development function that involves every business function. With data scientists, business decision-makers, and IT staff all working together in a collaborative framework, businesses in the APAC are seeing the opportunities that adopting agile AI offers. Learn how internal operations can pool data with external partners and customer metrics to drive a “virtuous circle” of insight and improvement by reading the full IDC InfoBrief. Getting a real competitive edge might only be a click away.
- Four Reasons why the next gen eCommerce platform is critical post-pandemic
- A real-time API platform for customers who demand real-time experiences
- How global firms are maintaining pace during unprecedented times with DocuSign
- Prevent, detect and respond – the security of Nuage’s SD-WAN
- Seize the opportunity to transform SAP and the enterprise, with SUSE