Tbk

Christian Angkasa, Enterprise Data Head, PT BFI Finance Indonesia

The AI Imperative in Financial Services

Christian Angkasa

AI and Data Reshaping Financial Decision-Making

The financial services industry has consistently pioneered the adoption of information technology, and the current era of rapid developments in AI and data analytics is no exception. I have observed that decision-makers within the sector are increasingly confident in the widespread integration of these technologies into their organizational frameworks and core decision-making processes.

The rapid advancement of AI capabilities has served as a critical wake-up call. It is clear that remaining passive is no longer a viable option. Evolving competition, shifting customer expectations and volatile global conditions require swift action and fundamental change.

We are currently at a convergence point where existing AI capabilities, external market momentum and internal improvement needs meet. This presents a significant opportunity for the industry to explore and implement adaptive solutions in several key areas:

• Hyper-personalized customer engagement: Delivering tailored experiences that meet individual client needs.

• Operational acceleration: Streamlining internal processes for greater efficiency.

• Advanced risk and fraud management: Enhancing security and oversight through predictive analytics.

• Knowledge management: Empowering field sales staff with reliable, real-time information.

• Productivity in IT development: Increasing the output and reliability of product development teams.

Key AI Use Cases Across Risk, User Experience and Fraud

There are already numerous real-world examples of AI use cases that enable financial institutions to identify and mitigate risks more effectively.

“We are currently at a convergence point where existing AI capabilities, external market momentum and internal improvement needs meet.”

AI models can simulate various market scenarios, analyze complex data and combine vast amounts of information that would be difficult or time-consuming to analyze manually or through traditional methods. From there, the AI model provides credit risk recommendations, identifies indications of vulnerabilities that can be further investigated and provides solution options with more detailed explanations following our commands, which we can input via prompts or voice commands. Isn't that sophisticated?

Use cases like credit risk assessment, fraud prediction and detection can be categorized as having a significant monetary impact. However, I see use cases in this industry that have a more fundamental impact, such as customer engagement using AI-based chatbots that can engage customers in natural, human-like conversations, providing instant assistance 24/7.

With the most advanced AI, these AI bots are not only rule-based; they understand context, sentiment and nuances in language, creating a very familiar and fluid dialogue, as if chatting with a real person.

Personalized marketing powered by AI can lead to higher customer satisfaction, increased cross-selling opportunities and a more significant return on marketing investments. Companies can deliver the right product or service to the right customer at the right time.

With the development of such capabilities, this can significantly increase customer handling capacity, enabling 24/7 operations, constant availability and consistent cross-selling and up-selling, as well as responding to the rapid resolutions customers expect.

What Differentiates Leaders in the Technology Race

In the current financial landscape, the gap between leaders and laggards isn't just about the size of the IT budget; it's about how technology is integrated into the core business. I see this not only as being outdated or missing trends, but also as being slow to adapt to this technology race. The impact is very serious for our organization's long-term survivability, even faster than our comfort zone.

At least the following are some important things that companies need to have in the era of the technology race, especially in the financial industry, which will determine the difference in the organization's survivability.

Implement and accelerate the development of data as a product: We should not just store data; we shall treat it as a high-quality product. It should be clean, well maintained real-time data pipelines that power very fast credit decisions and hyper-personalized customer experiences.

Agility and alignment are mandatory for the new way of working. Organizations that lead break down the silos between "The Business" and "IT." They use cross-functional teams where product managers and engineers share the same KPIs.

Leaders shall embed AI into deep-value areas such as automated underwriting, fraud detection, expanded productivity, and accelerated delivery, i.e., with AI-based code generation. For the customer-facing part, it can improve the customer experience with AI-based hyper-personalization.

Profitability and business continuity can be threatened by delays in adopting AI and being stuck in legacy systems.

I see adapting and capitalizing on this technology race as helping companies overcome classic technology issues such as ballooning technology operational costs that could be allocated to new product development, resulting in delays in responding to rapidly changing market needs; and slowing technical productivity due to the burden of legacy systems, which creates compounding "technology debt," drastically reducing competitiveness compared to companies that have already moved towards cloud-native and AI-first.

Leveraging AI Without Overcomplicating the Technology Stack

The most important and first step is strategic alignment rather than acquiring tools over tools. Then, fundamentally, effective adoption focuses on improving or upgrading existing data foundations and integrating AI into current workflows, rather than layering new, isolated software on top of legacy systems.

Specifically, in the lending or refinancing business, the application looks like this: revamp traditional credit scores with alternative datasets like utility payments and rental history to reach underserved borrowers; expand our market using advanced analytics with geospatial data; use AI, especially algorithms, to detect early warning signs of default by identifying patterns that conventional machine learning models created manually based on habits have previously missed.

We can start with a small POC (Proof of Concept), then prototyping before implementing it in more complex use cases. We can start on a small scale, for example, with a specific product or a few branch offices, before rolling it out nationwide. In the future, we can target integrating AI agents into a lending or refinancing workflow that will transform from a manual business process to an autonomous orchestration, even to the core business, such as embedding AI agents into the Loan Origination System (LOS) to accelerate decision-making and reduce operational complexity.

Challenges in Scaling AI with Governance, Compliance and Trust

Currently I see the biggest challenges that must be overcome are the people, followed by processes that comply with new regulations regarding technology. But I am very optimistic this challenge will be solved eventually as we are capable of adopting and see more incoming value that we deliver from it.

Meanwhile, the technology is not currently a significant obstacle; rather, it's a matter of company culture and human resource talent that can understand the implementation of both the transition of corporate culture with AI-assisted technology and synchronize with the strict GRC (Governance, Risk and Compliance) requirements of finance.

Financial institutions in Indonesia must ensure AI models—often hosted in global clouds—do not violate data sovereignty or cross-border transfer restrictions.

One of our initial achievements in this regard is formalizing and assigning a Data Protection Officer (DPO) within the organization and conducting Data Protection Impact Assessments (DPIAs) for potentially high-risk AI processing. To increase trust, maintain compliance, and ensure a level of trust that aligns with the industry's risk appetite, we can implement a hybrid cloud strategy, moving sensitive AI workloads from public clouds to private or hybrid environments to meet regulatory requirements for operational resilience.