Hardly a day goes by without AI profoundly affecting daily work, from predictive maintenance in the airline industry to real-time tracking of fleets and orders for logistics companies.
Across Asia-Pacific, the adoption of enterprise AI has surged in recent years. According to an IDC infobrief on AI sponsored by Daitaku, in 2021 and 2022, only 39% of businesses in this area integrated AI into their operations, but a remarkable 76% of enterprises have since embraced AI to fuel growth, improve efficiency, and spark innovation. This shift is underscored by the projected 40.8% increase in AI platform market spending in Southeast Asia from 2022 to 2026.
As organisations eagerly embrace the potential of AI to gain a competitive advantage, from better customer experiences to improved team productivity, there are potential obstacles that could obstruct the realisation of this value. A recent Daitaku survey titled “AI, Today” found that when it comes to enterprise AI adoption, data quality and speed of deployment were identified as top roadblocks. Conversely, leaders in AI considered cost and the lack of business use cases as less significant barriers to unlocking greater value from AI initiatives.
Given the rapid pace of digitalisation, overcoming these obstacles will be crucial to empowering organisations to derive more value from their AI investments and make it a scalable, sustainable answer to many business needs. This starts with recognising AI as a natural transition for maintaining a competitive edge or risk becoming a laggard.
Charting the course for AI success
Adopting AI demands extensive cooperation and collaboration across teams, yet the potential return on investment (ROI) is undeniably promising. According to PwC research, AI is projected to contribute US$15.7 trillion to global economic growth by 2030. The question then is, who will seize the lion’s share of the rewards? The answer lies with those who proactively take the lead.
Companies that succeed in harnessing AI technologies and putting them to work will enjoy a sustainable competitive advantage in their markets. Those who miss this opportunity will fall behind, potentially finding themselves in a position from which they cannot recover.
However, the road to AI adoption is complex and can be treacherous without the right guidance and roadmaps in place.
Here are some key challenges and solutions to consider when charting the course for AI success:
- Defining a clear strategy
The growing buzz around data science, machine learning, and AI puts mounting pressure on organisations to revamp their business models with advanced data solutions, or risk lagging in a rapidly evolving tech landscape. However, without a clear AI strategy and goals, organisations often miss the most valuable starting points for incorporating AI, ultimately leading to diminished value and trust.
But how do decision-makers select the right AI project? Should it be a consumer-oriented solution like a chatbot, virtual assistant, or computer vision application? Or would a focus on embedded, back-of-house efforts that quietly enhance internal operations, workforce efficiency, or decision-making be wiser?
Leadership teams should engage with all stakeholders, establishing effective communication to evaluate the business value, effort, and likelihood of success, strategically choosing and prioritising projects that can bring AI benefits to the organisation. If value cannot be proved, particularly in terms of ROI, further investment could potentially stall. It’s only with a solid foundation, comprising both top-down and bottom-up contributions, that a more comprehensive and robust AI vision can take shape. - Closing the talent gap
Insufficient skills and expertise often pose obstacles in the initial phases of AI adoption; when there isn’t a well-established data or AI team, addressing these shortcomings becomes daunting. A typical AI project requires a highly skilled team, including a data scientist, data engineer, and designer — and there simply aren’t enough skilled professionals available.
The rise of AI has also meant that workers require specialised skills and experience in areas such as natural language processing and robotic process automation, yet the talent pool remains scarce. In fact, the World Economic Forum has predicted that by 2025, up to 85 million current jobs may vanish, while 97 million new jobs will emerge, placing higher demand on technical skills. This shift is attributed to what they have termed “the new distribution of labour among humans, machines, and algorithms.”
AI adoption requires data scientists’ technical expertise, but building a scalable and sustainable data team involves more than finding an elusive “data unicorn” with all-encompassing skills. Leaders should facilitate collaboration between data scientists, engineers, domain experts, and business stakeholders, encouraging regular communication, establishing interdisciplinary teams, and promoting knowledge sharing. By investing in education and making data accessible to all, business analysts can be empowered to become citizen data scientists towards a well-rounded data team. This is where ‘everyday AI’ comes into play – a reference to AI that is intrinsic and seamlessly intertwined with the daily operations of a business, not solely used or developed by a central team. - Building a solid foundation
The challenge of AI adoption extends beyond model design; it hinges on the underlying technical infrastructure. Regardless of how well your AI model is designed, inadequate technical infrastructure compromises success. This architecture must serve three vital functions:
- Provide stability and reliability for existing users, ensuring smooth operation and high performance.
- Facilitate seamless onboarding and operation for current needs, making it easy to integrate new data sources and users.
- Be future-proofed for scalability as the organisation expands.
Investing in a data architecture that meets these requirements enables organisations to unlock the full potential of their AI initiatives. Though assessing the ROI of architectural upgrades can be challenging, collaboration between data teams and R&D to establish feasible goals, evolving as AI integration deepens, can overcome this hurdle.
- Governance and risk management
Crucially, governance, control, and trust are also integral to building a solid foundation for AI adoption. As more organisations experiment with and deploy AI in production, these applications need to be both safe and scalable in an enterprise context.
Internal stakeholders must remain in control even as they scale and regulatory compliance measures evolve. Moving towards systemising the use of data and AI for better day-to-day decisions requires organisations to invest in the right governance to address these challenges before it’s too late. - Breaking down data silos
A foundational aspect of AI adoption is fostering an organisation-wide data culture. For those working with data, encountering roadblocks due to siloed data — like constantly having to request access or not even knowing what data exists to work with — is frustrating and therefore delays or obstructs progress. This frustration incurs a cost: lost time, incomplete data projects, and incorrect models. The time-consuming struggle to access data or even identify its existence drains productivity — the larger the team and the more fragmented the data, the greater the financial toll of diverting personnel to data retrieval tasks.
More detrimental than an incomplete data project is an inaccurate one. Accessing siloed data, often lacking context, presents challenges. Without central oversight, teams may use incomprehensible data, misguiding the business and leading to flawed decisions and models.
To scale machine learning within an organisation, it’s necessary to thread together data, technology, and people, all in a governed manner. This thread is a centralised platform for users, both technical and non-technical, to maximise data-driven insights through the development and use of AI-powered applications.
The value of everyday AI
AI holds immense potential to deliver substantial value, and today, most organisations acknowledge the importance of integrating data and AI into their core operations for survival. However, the challenge lies in the intricate process of AI readiness and bringing initiatives to fruition, particularly in light of rapid developments in generative AI.
Facilitating quicker execution by involving more people in the analytics processes aims for AI integration so seamless that it becomes an intrinsic part of daily business operations, not confined to the use or development by a central team.
Empowering the workforce with AI potential doesn’t entail turning everyone into data scientists but rather making data science skills accessible to all within the organisation. Companies that equip their employees to grasp data, utilise user-friendly tools, and make AI-driven decisions will emerge as true winners.
