Maybank Malaysia: Redefining Customer-Centric Banking

In the competitive world of financial services, understanding and engaging with customers effectively is paramount. For Maybank Malaysia, the largest financial services group in Southeast Asia, achieving this goal required overcoming significant challenges:

• Fragmented customer data
• Inconsistent collaboration across departments
• Inefficiencies in data workflows

Recognizing the need for transformation, Maybank turned to Dataiku to drive innovation, scalability, and personalization in its customer engagement strategies.

A Two-Fold Challenge: Fragmented Data and Inefficient Processes

Maybank faced two key challenges. First, siloed customer data made it difficult to build a unified 360° view of customers. Teams spent valuable time exploring datasets and performing manual feature engineering tasks. With no standardized data repository or documentation, different departments often duplicated efforts, leading to inefficiency.

Second, the bank’s go-to-market process, crucial for delivering personalized recommendations, struggled with a lack of streamlined tools for multiple departments to collaborate – resulting in long turnaround times.

Breaking Down Silos With Feature Stores

To address these issues, Maybank implemented feature stores using Dataiku, transforming how the bank accessed and utilized data. Feature stores served as centralized repositories that consolidated customer data across business units, enabling teams to create a 360° customer view. This system was modular, allowing different departments to collaborate efficiently by leveraging pre-built and documented feature groups.

In terms of automation, Maybank used Dataiku scenarios to automate the pipelines, run campaign tracking, and send out csv files and PDFs of dashboards with results to relevant stakeholders in emails. On top of this, they built Dataiku applications to empower less sophisticated end users to be able to generate leads, engage in market sizing, and create dashboards, while being able to configure parameters as needed.

The impact was immediate and transformative. In January 2024, feature stores supported 44 projects, but by the end of the year, they powered over 700 projects. Maybank also increased the number of features stored from 2,400+ to 5,000+, creating a scalable foundation for data democratization. Teams could quickly iterate on workflows, adding new features to existing projects in as little as a day, reducing time-to-market and fostering a culture of reuse and collaboration.

Additionally, the automation of ETL pipelines and workflows enabled cleaner visualization and greater organization of workflows into zones. This resulted in less errors and more visibility of their data processes.

Dataiku has significantly accelerated our ability to translate data into tangible business value. The platform’s collaborative nature has streamlined our data processes, enabling us to develop and deploy data-driven solutions faster and more effectively than ever before. It’s a key enabler of our data strategy. Simon TG Lim Group Chief Data Officer at Maybank Malaysia

Scaling Personalization

Personalization is a cornerstone of customer-centric banking, and Maybank revamped its approach. Multiple machine learning (ML) models, feature stores and pipelines were built in a series of modularised parts across multiple departments. These were then combined to create custom engine(s) to address business objectives, using a  mix-and-match, plug and play approach which enabled fast builds of end-to-end data pipelines to power specific campaigns and business use cases – essentially a “Customer 360 Playbook,” using advanced ML and Dataiku’s automation capabilities. This saw a 10x increase in engagement rates on campaigns.

One of the key modules included a Dataiku-powered NBO model — a single model capable of serving multiple product lines. The model integrated data from customer transactions, demographics, and engagement histories to predict the most relevant offers for individual customers. The platform’s explainability features also allowed them to maintain transparency and compliance with internal regulations.

The results were outstanding:

• A 3x increase in product uptake among targeted customer segments.
• A 10x reduction in time and effort required for model maintenance, as a single scalable model replaced multiple siloed ones.

The NBO model empowered Maybank to deliver hyper-personalized recommendations dynamically while ensuring compliance through Dataiku’s governance features. As the model evolved, parts of multiple NBO models created by different departments were combined into one, further increasing inter-departmental collaboration.

With Dataiku’s automation and integration capabilities, we’ve drastically reduced the time spent on manual data processing and campaign execution. The model also allowed us to scale personalized offers efficiently, increasing customer engagement and satisfaction. Qamra Jema Khan Lead Data Scientist, Hyper Personalisation and Advanced Analytics at Maybank Malaysia

Accelerating Decision-Making and Collaboration

With Dataiku, Maybank built a data-first culture by breaking down silos and democratizing access to analytics tools. Data practitioners across skill levels — from expert data scientists to non-technical teams — collaborated seamlessly on projects. The platform’s automation capabilities reduced manual processing tasks and allowed teams to focus on strategy and innovation.

Dashboards and reports generated through Dataiku ensured transparency and real-time campaign tracking, helping stakeholders make informed decisions. Dataiku models examined the feedback loop, which allowed Maybank to determine the most important features and customer personas for a specific campaign context.

Self-service applications enabled less technical users to generate insights and size markets independently, fostering inclusivity and agility.

All these Dataiku features allow us to build modules and flexibility into our workflows in a collaborative way, allowing us to build fast, experiment fast and iterate fast. Qamra Jema Khan Lead Data Scientist, Hyper Personalisation and Advanced Analytics at Maybank Malaysia

A New Era of Customer-Centric Banking

Maybank’s transformation delivered measurable business value:

• Revenue increased by several million ringgit annually through enhanced customer engagement.
• The “build-once, reuse-many” approach saved time and IT resources, with workflows supporting rapid scalability.
• Go-to-market timelines for personalized campaigns were drastically reduced, enabling Maybank to stay ahead of the competition.
• Cross-departmental collaboration flourished, driving cultural change and embedding data-driven decision-making across the organization.

Dataiku as a platform has been instrumental in democratizing and unlocking data access and insights with empowering our teams at Maybank. We've seen incredible collaboration and upskilling across the organization. As we look to the future, Dataiku's robust platform is key to scaling our data initiatives, enabling us to unlock even greater value from our growing data assets and empower an expanding community of data users. Muhammad Haseeb Masud Qureshi Head, Group Data Platform & Architecture at Maybank Malaysia

Maybank’s journey with Dataiku underscores the potential of modern analytics to drive business success. By combining centralized data infrastructure with scalable ML models, the bank has not only optimized its operations but also set new standards for customer-centric banking. As Maybank continues to innovate, its partnership with Dataiku will remain a cornerstone of its mission to deliver superior value to customers while fostering financial inclusion across Southeast Asia.