How to hyper-personalise finance: Discover more about the project approved by CDTI

As I mentioned in my previous article, our main goal is to develop and implement new functionalities in the Coinscrap Finance’s financial coach. Thanks to this, our clients, including some of the leading banks globally, will be able to make strategic decisions based on the analysis of behavioral patterns.

Table of Contents

To achieve this, we will implement three new developments that will allow us to increase the added value of our module. 

Óscar Barba

Co-founder & CTO of Coinscrap Finance

These are the keys to better understand the banking user:

1. Generate user segments in real-time.

The concept or description of banking transactions is extremely useful for entities. Due to all the information it contains, a deep understanding of the needs and preferences of each user is achieved, which in turn facilitates the delivery of personalized recommendations and actions.

From the point of view of our clients, this will allow them to deeply analyze their consumer typologies, in addition to enabling a greater understanding of the needs and preferences of each person.

2. Improve the system’s ability to make banking predictions.

Based on the analysis of historical data, a model will be developed to identify financial patterns and trends, as well as make accurate predictions about future events.

This provides users with an advanced view of their financial situation, allowing them to make informed decisions and better plan their finances. The benefits of this functionality include greater control and planning capacity for users.

All of this translates into facilitating their financial well-being.

3. Offer a hyper personalized financial experience.

Through the analysis of historical data, the prediction system draws a series of conclusions. The system accompanies the user in monitoring their personal finances, providing recommendations, relevant insights, and alerts about potential financial products that may interest them.

This implies greater interaction and commitment with the brand, not to mention increased conversion to sales in available financial products and services.

Some key goals when improving a PFM or expense manager

Conducting thorough prior research on the state of the art in user segmentation is paramount when developing our project. We employ machine learning algorithms, personalized recommendation techniques based on artificial intelligence and deep learning.

This allows us to understand the latest trends and advances in the field and establish a solid foundation on which to work.


The scientific-technical analysis of unsupervised algorithmic models based on the mentioned technology is the next step for outstanding user segmentation. Different approaches and techniques are evaluated and compared, identifying the most suitable for this project. We then design and develop the real-time segmentation module, which allows us to design and develop the hyper personalized recommendation module.

The same process occurs with the forecasting module. We use time series analysis techniques and machine learning to predict users’ inclination to purchase products. We then proceed to integrate all developed modules into a coherent and functional solution.

This is the final phase of the project, in which exhaustive testing is carried out, in addition to documenting the entire technical process.

Take a look at our project “Hyperpersonalised Financial Intelligence

We delve into the technical characteristics of our project

“Hyper personalized Finance Intelligence” represents a significant advancement in the design of Personal Finance Management tools. Starting from a powerful bank movement categorizer, we can create a new system with the ability to detect user behavior patterns, as well as extract knowledge from their purchasing habits and product preferences.

Specifically, we aim to address the following aspects in a cross-cutting manner:

Clustering of users by similar movements.

Groups of people with similar financial and commercial preferences will be identified.

Identification of relevant events.

The system will identify events that constitute outliers in user behavior (unexpected income or expenses) and specific events of interest (real estate purchases, unusual high incomes such as inheritances, personal situations).

Product recommendation.

 Types of products (categories and products) acquired through bank transactions will be detected, and new products will be recommended to other users through movement similarity analysis.

Prediction of future product acquisition.

Analyzing movement patterns will predict the inclination to purchase products. The main difference with the product recommender is that this approach is based on pattern analysis in historical data prior to purchase.

Each of the objectives listed above will be addressed by combining supervised and unsupervised Machine Learning techniques. Supervised algorithms require labeled datasets (De Arriba-Perez, 2020). Behavior patterns will be extracted from these datasets using the existing mathematical relationships between the variables introduced into the system and the target category.

On the other hand, unsupervised systems will be used to detect clusters of inputs whose variables are related to each other and therefore are similar (Naeem, 2023) in the target domain. These same systems are capable of detecting elements that are far from the rest of the samples (isolated users or outliers).

ebook Cashback  EN

The lack of advanced technology and personalization in the traditional financial sector limits its ability to offer modern financial experiences tailored to the needs of new generations. At Coinscrap Finance, we are confident that this innovative project will fill this gap and enable entities to stand out and achieve levels of engagement and understanding of their customers never seen before.

About the Autor

Óscar Barba is co-founder and CTO of Coinscrap Finance. He is an expert Scrum Manager with more than 6 years of experience in the collection and semantic analysis of data in the financial sector, classification of bank transactions, deep learning applied to stock market sentiment analysis systems and the measurement of the carbon footprint associated with transactional data. 

With extensive experience in the banking and insurance sector, Óscar is finishing his PhD in Information Technology right now. He is an Engineer and Master in Computer Engineering from the University of Vigo and Master in Electronic Commerce from the University of Salamanca. In addition, Scrum Manager and Project Management Certificate from the CNTG, SOA Architecture and Web Services Certificate from the University of Salamanca. He recently obtained the ITIL Fundamentals certification, a recognition of good practices in IT service management.

Sign up for our newsletter and get our top stories delivered straight to your inbox