About the client:

A universal bank that has been operating in one of the CEE countries for about 20 years, serving retail, consumer finance, agricultural, small and medium-sized enterprises, and corporate banking.

Challenges

Due to dynamic growth and scale of operations, the bank identified several business challenges concerning the data architecture.

Regarding aggregated data, growing internal demand for:

Vector-1

real-time data for operational and analytical purpose

tailored-integration

better access to data

Vector Search - ai prompter In banking

new data sets for business users

multi-location

taking advantage of unstructured data

Additionally, new challenges with processing non-aggregated data and real-time events arised.

Challenges were also identified on the IT Department’s side:

project-management-support

increase of new data sources and volume of data

data-ai-ml

increase of infrastructure cost for data

x-icon-red

effective work with data exceeded the capabilities of existing legacy systems & IT technologies used at the time of evaluation

Next-generation Customer Data Platform to the rescue

Financial companies today face a rapidly growing volume and variety of data. Additionally, financial service decision-makers need support to meet

  • the challenges of high costs
  • the complexity of implementing and managing traditional databases
  • the growing competitive landscape

 

All this requires a next-generation data platform that

  • can adapt to meet the most demanding requirements
  • is cloud-ready
  • scales to meet growing business needs
  • is able to adjusts to both transactional and analytical workloads.

 

The bank decided that current data access by enterprise document management (EDM) doesn’t meet the requirements of a modern digital bank. After evaluation of solutions available on the market, the bank had chosen to build a Customer Data Platform (CDP) with Ailleron.

Perspective laptop position with business analysis app dashboard. Analysis of readings and data, online statistics, business forecast and progress tracking. Business dashboard on laptop. Vector

Ailleron solution for a next-generation data platform

 

Ailleron Customer Data Platform (ACDP) is a set of tools and technologies that create a comprehensive solution for processing and managing data from various systems.

 

Our offer was accepted due to:

  • flexibility concerning the supported technologies and alignment with the bank’s corporate technology stack
  • detailed architecture of ACDP with a clear vision of solving bank’s business & IT challenges
  • readiness for a quick start – in only 2 weeks from bank’s approval decision
  • availability of experienced specialists – we employ both software development and financial business experts with +10 years of hands on experience & domain background

Main ACDP use cases:

Ailleron CDP Challenges
  • Providing data for Personal Finance Manager (part of client’s Digital Banking Platform)
  • Providing data for 360 Customer Views

 

 

Scope of data processed by ACDP includes:

 

  • All customers’ transactions with categories
  • Customer personal data
  • Products

In this case, ACDP development is based on bank’s data centric architecture principles:

start

CDP as the primary source of data provided across the whole organization in a specified country

start

Data Stream Processing as the main data flow from source systems

start

Data governance and organization based on Data Mesh approach

covering Data Producers, Consumers, and CDP

Main components of ACDP provided during the project

tailored-integration

Data Streaming

Core component of ACDP responsible for:
– gathering events and data from source systems (events, CDC Connectors, etc.)
– distribution of events within the whole platform
– data transformations

 

data-science

Data Lakes

Storing data from various sources for analytics purposes:
– data lake with raw data
– data lake with enhanced data

multi-location

Operational Data Stores (ODS)

ODS are used to store optimized data for external systems (like Internet / mobile banking, CRM, back office applications). In this case, they store domain data like:
– transactions
– products
– customers
– etc.

Vector-1

GraphQL

Provides read-only API and is responsible for aggregation of data from 1 or more ODSes for particular customer views

custom-software-services

API Gateway

– Takes all API calls from clients and routes them to the appropriate microservice
– Provides security and access control

banking-business

Data Hub

Component responsible for data discovery, data governance, metadata management, and data lineage; allows for
– adding tags to data
– describing data attributes

e-leasing

GDPR

Component responsible for data anonymization (when it is required by bank)

Implementation

Data Analytics, Computer Screen And Woman In Night For Stock Market Research, Graphs And Chart Analysis. Trading App, Digital Overlay And Software Information Technology Of Business Person Thinking

The project is being carried out in

  • agile methodology
  • strict cooperation with the bank, including daily meetings & statuses

 

Crucial roles among Ailleron Team include:

  • Data Platform Architects
  • Data Engineers
  • Data Platform Developers
  • Data Analysts

 

Ailleron Team scaled up from 4 FTE to 12 FTE in just 3 months.

Implementation implies ACDP delivery in a step-by-step process

Project kick-off included some of the following steps:

1. Preparing backlog for next sprints

2. Implementation of ACDP core

data streaming platform, data lakes, ODS infrastructure

3. Integration of the first data source

4. Preparing the complete data flow

5. Measuring the whole data process with KPIs & reconciliations

6. Adding new data sources

7. Etc.

Key parameters (KPI) achieved by ACDP in the project

awards

3 seconds to deliver data in Data Lake and ODS

awards

50 milliseconds to make data available for external systems like PFM, 360 view, etc.

awards

~10 billion records of transactions stored in ACDP

Vector-1

12 defined data sets shared for external systems

ACDP tech stack used in the project

24e03322-742a-4515-893b-62b0f872d553
  • MongoDB
  • Kafka Confluent
  • Apache Spark
  • Prometheus
  • Graphana
  • Delta Lake
  • OpenSearch
  • Zipkin
  • Vault
  • Fluent ID
  • kStream
  • Aqhq
  • Oracle
  • Java
  • PySpark
  • ksql

Mastering data management gives our client that edge to become the first choice institution for customers looking for future-ready banking. This is a great privilege to see how Customer Data Platform powers highly innovative financial services in a mature market. Seeing real-time data sharing for advanced analytics and key systems, such as mobile banking, online banking, and CRM makes the whole project team proud!

Michał Walerowski

Business Unit Director AI/ML & Data Solutions

abstract lines

Let’s make financial experiences
easy and enjoyable together!

Tell us what you need and we will contact you shortly.

Tell us what you need and we will contact you shortly.