Latest News & Insights

The Major Challenge and AI Application in Organizations

The major challenge for almost every organization right now is how to apply **AI** to generate the greatest benefit possible. The driving forces, whether from the board/executive level or from the demands of operational staff, are all aimed at increasing revenue, reducing costs, and boosting work efficiency through the use of AI. One critical question is: what plan or strategy should we have to successfully implement AI within the organization?

What is AI Strategy?

AI Strategy is defining the direction for AI application in an organization, including managing AI-related risks, to support the organization's core mission. This AI strategy must align with the data strategy and support the overall organizational strategy, with **AI Governance and Data Governance serving as the fundamental foundation**. At this point, one might ask why data is so important to AI. The simple answer is that for AI to truly meet organizational needs, it must undergo a **data training** process and requires the use of the organization's quality data to help train the AI. And in sub-fields of AI, such as Machine Learning or Deep Learning, learning from data is essential. Simply put, if we want high-quality AI that meets the organization's needs, we must let it learn from the quality data within our organization.

Why is an AI Strategy Necessary? Beware of Data Leading to AI Silos

Driving AI initiatives is no different from driving other initiatives. Without a strategy, there is a high chance that projects will lack connection and alignment with the organization's goals, resulting in outcomes that fall short of expectations, and an inefficient use of capital investment, personnel, and time. Given the rapid pace of change in AI technology, we don't want to use AI use cases that are quickly becoming outdated, can't be developed further for the future, or use cases that can't operate autonomously (automate).

Many organizations today that want to push AI initiatives have already started, which shows a positive sense of urgency. But in the big picture, we must discuss where we go next. If every unit acts independently, there is a high chance of seeing scenarios where sub-units within the organization use various forms of AI use cases. The risk to be aware of is that we might have a proliferation of AI use cases that end up as **AI silos**—meaning everyone acts independently, resulting in duplicated investment, inefficient resource utilization, and ultimately, a lack of alignment with organizational goals.

I believe **AI silos** are something organizations want to avoid because they mean a lack of integration and systematization in usage. And from a budget perspective, since the cost of using AI is relatively high, AI silos mean significantly increased costs for the organization, while departments lack coordination (Synergy) and fail to achieve Economies of Scale.

The Role of AI Strategy in Supporting Organizational Strategy

I believe one of the key questions for everyone in organizations right now is whether their current strategy remains effective in an era where AI is set to change the behavior and patterns of most people, requiring organizations to adapt. In this context, developing an **AI strategy helps to review and revise the organizational strategy to stay current with the near future**. A strategy remains relevant only if it enables the organization to adapt before it is harmed or destroyed. The saying, **"Disrupt before being disrupted,"** still holds true in every era of change.

Furthermore, we may need to accept pain in certain parts of the organization and at certain times in order to adapt the business model to the AI era. In fact, accepting partial pain for the sake of change is not a new concept; if we remember, there was a time when **Apple** chose to discontinue the iPod to allow the iPhone to emerge.

Another example is **Google**, which, despite generating massive revenue from search engines and Google Ads (58% of Google's 2022 revenue came from Google Ads), chose to embrace the arrival of AI and develop **Gemini**, a Multimodal LLM, to grow with the AI trend in the future.

A third example is **Disney**. After allowing Netflix to be the king of the Streaming subscription business for many years, Disney made a major pivot by jumping into this sector with Disney+, Star+, and Hotstar (as of 2024 data, although the Disney group still has fewer subscribers—approx. 200 million—than Netflix—approx. 300 million—it shows a full-scale adaptation to grow with the AI trend).

AI can create new business opportunities and offer more detailed choices and channels for consumers. It is now widely accepted that businesses in any industry can effectively create pricing strategies, understand customer behavior, target specific customer groups, and establish new business strategies to increase competitiveness with the help of AI.

However, challenges remain in terms of technology, infrastructure, and finding the relevant data to create AI that meets organizational needs while using data ethically and responsibly (governance).

Why Must Data Strategy Go Hand-in-Hand with AI Strategy?

In the last article, I discussed that the key factors driving AI transformation are vastly more powerful processing systems and the massive amount of data being created and stored. However, having a large volume of data alone does not guarantee that we will get knowledgeable AI that meets organizational needs. **Data quality is a crucial factor** that ensures the AI initiative truly addresses our organization's challenges. This is why data strategy and governance are incredibly important in driving AI initiatives.

According to a study by Scale AI, leading technology companies worldwide invest in AI with the following breakdown: **60% for processing systems, 30% for data**, and 10% for algorithms. This figure reflects that driving AI initiatives is not just about focusing on hardware to increase processing power or focusing on coding without incorporating quality data for training.

Regarding Data Governance itself, having quality, accessible data is fundamental to building and training accurate and reliable AI models. Furthermore, the data used must be secured and treated fairly. Good data management helps ensure that data is used within ethical and security frameworks, and in compliance with standards and laws. When good Data and AI Governance are in place, leveraging AI has a higher chance of creating business value for the organization and correctly and appropriately managing AI Risk (AI Risk Management).

AI Use Cases Must Be Scalable for Efficient and Cost-Effective Implementation

Successful AI use cases, or **AI usecases, must be scalable** to grow from small pilot projects into large-scale enterprise solutions. **Scalability** ensures the system can handle increasing data volumes, users, and workloads without sacrificing performance, accuracy, or cost-effectiveness. Without a scalable design, AI projects often fail when faced with real-world demands. The details on why we need scalable AI are as follows:

Business growth and evolving needs

• Support More Users: As a business grows, AI applications must support a larger customer base or more internal staff. For example, an e-commerce recommendation system must scale to deliver personalized content to millions of users without latency.
• Handle Increased Data Volume: AI systems require large amounts of data. With exponential data growth, scalable systems are essential to efficiently ingest, store, and analyze increasingly larger datasets for training and inference.
• Enable New Use Cases: A scalable AI architecture is flexible enough to be adapted for new purposes without the costly need for a complete rebuild. A single, powerful AI system can be extended across multiple departments for different tasks, such as enhancing customer service, marketing, or finance.
• Ensure Long-Term Viability: Scalable AI use cases are future-proof. They are designed to adapt to new technologies and changing business requirements, generating a long-term return on investment.

Cost and efficiency optimization

• Maximizes ROI: Investing in scalable AI prevents expensive and time-consuming redesigns that non-scalable systems would require when they hit capacity limits.
• Optimize Resource Allocation: Scalable systems dynamically allocate processing resources, using only what is necessary at the moment. This flexibility reduces over-provisioning and lowers infrastructure costs.
• Reduce Operational Inefficiencies: Non-scalable AI systems often suffer from delays, performance bottlenecks, and high maintenance costs. Scalable processes adapt dynamically to demand, offering high flexibility and ensuring smooth operations.

Competitive advantage

• Elevate Customer Experience: Scalable AI can deliver a consistent, high-quality personalized experience for a growing user base. For example, Netflix uses a scalable recommendation engine to tailor content for hundreds of millions of members.
• Support Better Decision-Making: With more data, scalable AI systems can continuously improve insights, allowing the business to make more accurate, data-driven decisions.
• Improves Time to Market: A scalable AI framework enables the rapid deployment of new models and features, allowing organizations to quickly launch new capabilities to stay ahead of the competition.