- Who: Discovery Bank in partnership with Databricks.
- What: Implementation of a hyper-personalization framework using behavioral AI and governed data.
- Where: Operating in South Africa.
- Key Technology: AI-driven infrastructure for real-time decisioning and "reusable" data products.
- Primary Goal: Transitioning from demographic-based marketing to behavior-driven client interactions at scale.
Discovery Bank, a digital-first financial institution based in South Africa, has detailed its strategic implementation of behavioral AI and governed data to deliver hyper-personalized banking services at scale. By partnering with Databricks, the bank has moved away from traditional demographic segmentation in favor of a real-time decisioning engine that responds to specific client behaviors. This infrastructure allows the bank to maintain the speed and security required for modern financial services while ensuring every client interaction is tailored to individual needs.
Defining Hyper-Personalization Through Behavior
In a recent industry report, Discovery Bank defined hyper-personalization as the ability to make every client interaction relevant to a specific person at a specific moment. According to the bank, this relevance is determined by actual behavior rather than broad demographic categories.
The bank contrasts traditional methods—such as sending a generic savings prompt to every customer under the age of 35—with its new behavioral approach. For instance, the system can now identify a client who has recently received a salary payment, possesses a maturing fixed deposit, and has visited the investment section of the mobile app multiple times within a single week. By identifying these specific signals, the bank can surface a highly relevant suggestion in real-time, a feat that would be impossible with manual configuration at the scale of a national financial institution.
Building the Bank Around Data Products
Since its launch in 2019, Discovery Bank has aimed to transform the South African banking landscape by building its core operations around data products. According to Stuart Emslie, Head of Data at Discovery Bank, the institution’s foundation is rooted in behavioral science and AI. Instead of deploying isolated AI models for specific tasks, the bank embeds data products across its entire business ecosystem.
This integrated approach processes a continuous stream of signals generated by the client’s relationship with the bank. These signals include:
- Payment and spending patterns.
- Savings habits and fixed deposit maturities.
- Digital engagement and app browsing history.
- Borrowing decisions and service-related conversations.
The primary technical challenge identified by Discovery Bank is converting these disparate signals into actionable "next steps" while maintaining a seamless connection between personalization, fraud protection, and regulatory governance.
The Role of Governed AI Infrastructure
To manage these complex data streams, Discovery Bank utilizes Databricks to maintain a governed AI infrastructure. The bank emphasizes that as systems move from generating insights to taking automated actions, governance must remain present.
A key component of this architecture is "reusable decisioning." This allows the bank to maintain consistent definitions and controls across various departments, including marketing, digital experience, and servicing. By using reusable analytical products and deterministic services, the bank ensures that a client's "financial wellness" status or risk profile is interpreted identically whether it is being used for a marketing campaign or a fraud prevention check.
This unified infrastructure currently supports several critical business functions:
- Financial Wellness: Tailored advice based on spending and saving habits.
- Personalized Journeys: Dynamic app experiences that change based on user intent.
- Fraud Protection: Real-time monitoring of spending patterns to detect anomalies.
- Banker Assistance: Equipping human staff with AI-driven insights for better service.
- Generative AI and Agents: Implementing controlled actions through autonomous or semi-autonomous AI systems.
Impact on Financial Services
The collaboration between Discovery Bank and Databricks serves as a practical blueprint for the broader financial services industry. The bank's experience suggests that AI provides the most value when it is not treated as a standalone tool but is instead connected to trusted data and deterministic services.
For developers and industry leaders, the primary implication is the shift from "insight" to "action." Discovery Bank’s model demonstrates that personalization is no longer just about understanding the customer; it is about having the infrastructure to act on that understanding in real-time without compromising governance or security. This requires a transition from manual rules-based systems to automated, governed AI that can handle the massive volume of signals generated by modern digital banking.
What's Next
Discovery Bank continues to expand its use of behavioral AI and data products. Stuart Emslie, Head of Data, has shared further technical insights into the partnership with Databricks via an official video case study, detailing how the bank maintains its "behavioral science" edge. No specific timeline for new feature rollouts was provided in the current announcement, but the bank remains focused on embedding generative AI and autonomous agents deeper into its service model.