A customer opens their banking app and asks about a personal loan. The bank’s propensity model ranks them as a strong candidate. There’s just one problem: their financial position has changed since the model was scored. For Indonesian banks building personalised services around a data lake, this is a practical question. How do you combine what you know about a customer with what is happening in their account right now?
The lake remains valuable for analysing behaviour and identifying opportunities. But where operational updates arrive in batches, an application may need information that hasn’t reached it yet. Let’s look at what that means in a customer conversation.
Balancing customer needs with compliance
Let’s continue the scenario of the customer asking a mobile banking assistant about a personal loan. A propensity model, scored in the data lake several days earlier, ranks this customer highly. On this score alone, the customer appears to be a strong candidate for a loan offer, pending eligibility checks.
However, in the core banking system, a live check of the customer’s records shows a substantially lower balance and an account in arrears, caused by transactions which occurred overnight. This may make the customer ineligible for a loan according to bank policy. The problem is that the data lake does not have this latest information, as it receives data in periodic batch jobs. With just the data lake, the bank would have made a loan offer to this customer, which it would later have to withdraw.
With the Denodo Platform, information is federated live from the data lake and the marketing system, as well as current core banking data through a governed, read-only API. The application makes one request, without first having to copy the operational data into the lake.
Since Denodo federates the current, single source of truth; the assistant withholds the offer and directs the customer to a representative for further investigation. The bank is saved both embarrassment and a potential penalty.
Building on the Data Lake
This approach gives the data lake a broader role in the customer experience. Its insights can be combined with operational information at the moment an application needs them.
Denodo queries the relevant sources and presents the result through a single logical access point. Teams can discover and request access to it through the Denodo Data Marketplace, then reuse it across customer channels.
What Makes This Work
Four capabilities underpin the approach:
Universal connectivity. The lake, core banking interfaces, SaaS applications and APIs can contribute to one customer view while remaining in their existing environments.
Zero-copy delivery. Live queries retrieve the operational information needed for the decision without requiring another persistent copy. Read-only access prevents changes to core banking records; query and workload controls still matter for production performance.
Governed access. Fine-grained policies and masking govern what each application or team can see through Denodo. Lineage and access records help teams understand where information came from and how it was used.
Semantic trust. Shared business definitions give applications and AI agents a consistent understanding of customer information across systems.
A Local Example and a Wider Business Case
The Indonesia Stock Exchange selected Denodo as a strategic component of its data architecture to improve market data distribution. Its work included combining real-time messaging and streaming data with API sources. This is another relevant local example of bringing distributed information together for timely delivery.
Bring Current Context into the Conversation
A useful personalised experience depends on knowing when an offer makes sense for the customer. Combining lake-based insights with live operational checks helps the bank make that judgement while the conversation is still happening.
Book a demonstration to see how Denodo can bring that context to your existing data architecture.
- Complementing Your Data Lake with Real-Time Personalisation in Banking - September 25, 2026