Blogs

Attribute to: Mr. Kayton Wan, Software Sales Leader, IBM China/Hong Kong Limited

Artificial intelligence, particularly generative AI, is a hot topic worldwide. Amid all the buzz, a question emerges: Is your company’s data truly AI-ready? Let’s reflect on the connection between AI applications and the readiness of data, and how data can be primed through scaling AI across your organization.

Indeed, data readiness for AI is more than just a buzzword; it embodies the preparation and organization of data to mesh with AI algorithms, ensuring optimal effectiveness in real-world applications. This process is not just about collecting data but also structuring it meticulously, labeling it accurately, and organizing it in a manner that best serves AI models.

At the IBM Hong Kong Technology Forum – the panel discussion of "Is Your Data AI-Ready?", experts from Alpha Square Technology; Hang Seng Bank; Data Literacy Association; and Hong Kong Shue Yan University shed light on the importance of data readiness, how to identify data suitable for AI, and the critical role of nurturing a data-centric culture in comprehending and embracing data readiness in the era of AI.

Infrastructure, Algorithm, and Data Readiness for AI Success

Mr. Arthur Wong, Founder & CEO of Alpha Square Technology, has more than 30 years of extensive experiences driving IT transformation with new digital technologies and solutions. He highlighted three crucial elements for successful AI implementation: underlying infrastructure, the AI model or algorithm, and data readiness.

While resources are commonly allocated to infrastructure and AI models, data readiness is frequently neglected. Data readiness is portrayed as a challenging, yet vital aspect that sets organizations apart, as it is intricately tied to industry-specific needs and internal processes. Drawing from his experience in the banking sector, he focused on the complexity of tasks like metadata management and profit allocation, which require substantial time and effort to resolve. While more data enhances AI accuracy, ensuring data quality is equally vital. He recommends that businesses should increase their allocation of resources and manpower to address such critical data-related challenges effectively.

Arthur also discussed challenges in data management, including scattered data signals within organizations, the lack of clear data governance policies in small and mid-sized organizations, and the cultural aspect of data utilization. He emphasized the need to consolidate disparate data sources into centralized repositories, manage data elements like metadata and lineage, for effective decision-making and AI modeling, and improving the systematic use of data for critical decision-making.

Furthermore, he zeroed in on working with unstructured data, citing his firm's experience in developing a messaging platform for WhatsApp and WeChat. Extracting valuable insights from unstructured data poses complexities due to factors like language diversity, audio content, and emojis in customer interactions. To address these, advanced AI models like Natural Language Processing (NLP), audio recognition, and image recognition are essential for effective AI applications, he opined.

Key Control Indicators for Data Readiness

Mr. Edwin Hui, Head of Data & Analytics Office at Hang Seng Bank, is a seasoned data professional who has been in the industry for more than 20 years. He now leads the bank’s data and analytics team. Edwin questioned the necessity of striving for absolute accuracy over speed, pondering whether a slightly lower accuracy, like 90%, can expedite processes, especially in marketing campaigns where efficiency is critical.

This point led to the discussion of Return on Investment (ROI), particularly relevant in heavily regulated sectors like banking. In this context, various Key Control Indicators (KCIs) gauge the efficacy of data readiness across diverse perspectives. Edwin elaborated on the intricate nature of these KCI layers, encompassing assessments of data accuracy and usability throughout the interconnected upstream and downstream systems. He reflected on maintaining strict data accuracy while considering the constraints of time and resources, ultimately raising questions about the necessity of striving for absolute perfection, versus, finding a balanced approach from an ROI perspective.

Moreover, Edwin acknowledged the difficulty in justifying data governance expenditures to business leaders, especially when looked at from the perspective of considering ROI. He stressed the significance of maintaining a long-term perspective and continuous efforts for data governance, data controls, and readiness, underscoring their role in contributing to the organization's success for business stakeholders who are essential in funding and benefiting from these initiatives.

Recognizing the value of a holistic data management approach, Edwin cited key pillars being implemented at institutions like Hang Seng Bank. He accentuated the need for establishing controls to govern data effectively, implementing data literacy programs to enhance the understanding of data's potential value as well as fostering analytics capabilities to unlock insights and drive decision-making. These components work in tandem, in favor of a holistic strategy that integrates governance, literacy, and analytics.

Data Literacy: A Critical Factor for ROI on AI Investment

Dr. Toa Charm, the Founding Chairman of the Data Literacy Association, brings a wealth of experience from his senior management roles at Cyberport, HSBC, IBM, and Oracle. He emphasized the importance of data literacy in achieving a return on investment (ROI) from data and AI initiatives.

Dr. Charm highlighted that data literacy is essential for organizations aiming to maximize ROI from their data and AI projects. This involves assessing the data culture within the organization and the data mindset of its employees. Simply having a large volume of data does not guarantee optimal returns.

Addressing challenges in metadata management, Dr. Charm underscored the need for clear data definitions across departments to foster collaboration and solve complex issues. The ultimate goal is to monetize data assets by democratizing data literacy throughout the company, enabling employees to use data for problem-solving and value creation.

He also stressed the importance of evaluating data intended for data lakes to effectively address business challenges and meet promised use cases. He pointed out the need to prioritize tasks such as data cleaning, preparation, and governance to focus on actions that add value to the company.

A feedback loop is crucial for successful data monetization. Dr. Charm noted that data and AI initiatives require a collaborative effort involving multiple departments, not just IT and data experts. To succeed, organizations must cultivate a data-driven culture and mindset, fostering data literacy among all employees to align their goals, language, and understanding for effective collaboration towards shared success.

Data Readiness in Education and Research

Dr. Connie Yuen, Associate Professor and Head of the Department of Applied Data Science at Hong Kong Shue Yan University, also serves as the Founding Director of the Big Data Laboratory. She shone a light on the critical role of data readiness for AI within a research and education context.

Drawing from a project focused on developing an intelligence system for early dyslexia identification in collaboration with NGOs, Dr. Connie discussed the challenges they encountered in data collection. These challenges stemmed from the lack of raw data retention by early education centers, necessitating substantial effort and time to gather necessary data. She emphasised the imperative to bolster data readiness across business and NGO sectors.

Citing real-world references, she shared about initiatives by Hong Kong universities to advance data literacy through industry-tailored credit and non-credit courses, shaped with input from industry partners. She also weighed in on the value of practical exposure, such as summer internships in industrial projects, to deepen students' comprehension of data readiness and literacy. Dr. Connie said she anticipates students to increasingly grasp and further understand these concepts over the years.

Achieving Data Readiness: The Foundation for Successful AI Implementation

The experts reinforced how pivotal achieving data readiness is to unlock the full potential of AI and scale for successful implementation.

Businesses need AI that is trustworthy, scalable, and adaptable. To help businesses capitalize on their AI opportunities, IBM offers watsonx – a data and AI platform that includes three components and a set of AI assistants designed to help them scale and accelerate the impact of AI with trusted data: ⁱ

• watsonx.data: Data lakehouse to manage and optimize data workloads, with open, hybrid and governed architecture for scalable analytics and AI. It simplifies complex data landscapes, eliminates data siloes, optimizes growing data workloads for price-performance, and manages as well as prepares data to improve the relevance and precision of AI. 
• watsonx.ai: Bringing together new generative AI capabilities powered by foundation models and traditional machine learning into a powerful studio spanning the AI lifecycle; tuning and guiding models with enterprises’ data to meet their needs with easy-to-use tools for building and refining performant prompts. With watsonx.ai, businesses can build AI applications in a fraction of the time and with a fraction of the data.
• watsonx.governance: Integrated platform to direct, manage and monitor AI activities, accelerating responsibility, transparency and explainability

While AI infrastructure and model development are paramount, data readiness – encompassing data governance, data literacy, and advanced analytics – is equally vital. Organizations must adopt a holistic approach, consolidating disparate data sources, establishing robust controls, and nurturing a data-centric cultureⁱⁱ where employees at all levels are empowered to understand and leverage data effectively.

In fact, 79% of organizations state that looking ahead twelve months, data will be more pertinent to their organization’s decision-making.ⁱⁱⁱ IBM offers insights on achieving data readiness, emphasizing data preparation, data cleansing, and data utilization.

To lead an organization with AI-powered, data-driven decisions, data literacy is a competency everyone needs, not just data scientists. The ability to understand, interpret and communicate using data with it is a significant skill for all employees. The four foundations of a data-literate culture are^iv:

• Democratizing data access across your enterprise: By creating a central, governed data repository, such as a data fabric, people across the organization can easily access and analyze data, enabling technologies like analytics and AI to improve workflows. IBM suggests companies adopt Data Fabric architecture to simplify data integration, governance, observability, lineage and master data management. To deliver quality data for generative AI, IBM Data Fabric provides organizations with a trusted data foundation, enabling businesses to leverage automation for data discovery, enrichment and protection with IBM’s data governance and quality capabilities, employing various data integration styles to deliver reliable data for AI workflows. This architecture is composable, allowing IBM to meet clients wherever they are in their data journey. ^v
• Organizing information in a clear and transparent manner: Establishing a platform for governed data access is just the first step. It's critical to then help all stakeholders - from technical to non-technical users - understand data's value, origin, and quality. Transparent, explainable data processes inspire trust in AI initiatives by making data lineage and flows clear. While not everyone needs data science expertise, all should understand data, its lineage and how it flows within end-to-end processes - not just one part - to search for data, access relevant data, and enable business applications with it.
• Empowering data citizens for responsible data use and actionable insights with AI: Training in data literacy is essential for enabling teams to use data and AI-driven insights responsibly, leading to better decision-making and enhanced business outcomes. By fostering a deep understanding of data tools and the ability to transform data into impactful, visual narratives, organizations can turn raw data into actionable intelligence. An effective data literacy program empowers employees to leverage data as a strategic asset, ultimately linking data insights to tangible business results.
• Cultivating empathy to nurture data advocates: Curiosity is vital for fostering a data-literate culture. Employees and leaders should challenge assumptions and validate that AI-generated recommendations align with the organization’s goals. It's crucial to identify the data literacy skills necessary for driving business success and to empower data advocates to deploy data and AI capabilities across different functions. This approach helps build a culture of data stewardship and fosters a network of data champions dedicated to continuous learning.

Ultimately, data readiness is a journey that requires commitment to data management and a data-driven mindset. By prioritizing it, companies can unlock the true power of AI and make informed, impactful decisions that fuel their growth and competitive advantage.

We’d like to thank our four guest speakers for their participation and their esteemed insights. They have enriched our discussions and have made the participants further understand the power of data readiness to successfully implement AI.

Mr. Arthur Wong (Left 1st in the photo above), Founder & CEO, Alpha Square Technology; Dr. Connie Yuen (Left 2nd), Associate Professor and Head, Department of Applied Data Science, Hong Kong Shue Yan University; Mr. Edwin Hui (Left 3rd), Head of Data & Analytics Office, Hang Seng Bank; and Dr. Toa Charm (Left 5th), Founding Chairman, Data Literacy Association attended the Panel Discussion of “Is Your Data AI-Ready?” hosted by me (Left 6th), as Software Sales Leader of IBM China/Hong Kong. The four guest speakers joined Mimi Poon (Left 4th), General Manager, IBM Hong Kong and me for a group photo after the panel discussion.

ⁱ Source: https://www.ibm.com/watsonx
ⁱⁱ Source: https://www.ibm.com/resources/the-data-differentiator/data-literacy
ⁱⁱⁱ Source:  Voice of the Enterprise: Data & Analytics, Data-Driven Practices, 451 Research, 2022 , https://www.ibm.com/data-fabric?lnk=flatitem
^iv Source: https://www.ibm.com/resources/the-data-differentiator/data-literacy
^vi Source: https://www.ibm.com/data-fabric?lnk=flatitem

About the author:

Kayton Wan, Software Sales Leader, IBM China/Hong Kong Limited

@LinkedIn

Kayton Wan is currently Software Sales Leader of IBM China/Hong Kong Limited. He leads a team of professionals who deliver innovative and tailored software solutions to clients across various industries and regions. With over 10 years of experience in the information technology and services sector, Kayton has a proven track record of driving revenue growth, customer satisfaction, and market penetration for IBM's technology portfolio.  His core competencies include sales, customer relationship management, IT strategy, professional services, and data center management.  Leveraging his accounting background and business acumen, Kayton understands the clients’ needs and challenges and provide them with optimal and cost-effective solutions. His goal is to help IBM in Hong Kong to achieve its vision of being the trusted and valued technology partner of enterprises.