Recommender systems for banking and financial services


Knowing how time-consuming traditional banking approaches can be, the banking and financial services industry is constantly looking for new ways to streamline their operations and improve their customers’ experiences.  

One innovation that has gained significant traction recently is the use of recommender systems. There’s no room for inertia and inefficiency in this industry, so introducing recommender systems to enhance communication between banks and their clients seems like a great move forward.  

Read on as we discuss the role and benefits of recommenders in the banking and financial services industry and explain how to use them in this context.  

What is a recommender system? 

Before we go on, let’s define recommender systems 

As you probably already know, recommender systems are used to provide the best recommendation of our product that would interest the client most (system output), based on the user data (system input).  

In the context of banking and financial services, they analyze vast amounts of data to offer tailored suggestions to clients. Data includes:  

  • User behavior (e.g. channel interactions, credit bureau reports, and product equipment),  
  • Transaction history (e.g. card, online, abroad, and currencies), and  
  • Demographic and other valuable info from KYC (e.g. location, education, age, and employment) 

Considering that online communication between clients and banks tends to be a bit chaotic, banks can use recommenders to understand how to get the best possible outputs from the system, relying on the inputs. These systems can be classified according to: 

  • System input,  
  • System output or 
  • Based on algorithms that run in the background and create the recommendation.  

The first two classifications, developed by Things Solver and rooted in practical experience, provide a nuanced understanding of recommender systems. 

In the classification based on system input, we can find inputs from: 

  1. Online datasets, e.g., user visits to the website,  
  1. Offline datasets, e.g., user data kept in the bank’s database, or  
  1. A mix of both datasets.  

In the classification based on the system output, the recommender can propose products the user already experienced or those that he might be interested in but still does not have; or a mixture of both kinds. 

Benefits of recommender systems for banking and financial services  

Although banks might be reluctant or unsure about implementing recommenders, let us reassure you. From our point of view, the benefits of using recommender systems in the banking sector are twofold. 

First and probably more important — let’s look at the benefits of recommender systems for customers.  

  • Personalized product recommendations: Based on the customers’ financial goals and preferences, recommenders can create tailored suggestions for everyone. Besides making them happy, this can lead to better decision-making. You should avoid jeopardizing customer satisfaction with ineffective commercial pressure by offering products that do not fit customer needs.   
  • Improved financial literacy and awareness: By showing relevant products and services, recommender systems can help educate customers about available options and promote financial literacy. 

On the other hand, the benefits of recommender systems for banks and other financial institutions are reflected in the following:  

  • Better customer satisfaction and retention: By creating and offering personalized recommendations, banks can truly ensure their customers’ satisfaction and loyalty. 
  • Improved sales and cross-selling: Thanks to recommender systems, banks can boost their sales rates and revenue by crafting targeted cross-selling of products and services.  
  • Data-driven insights: Analyzing user behavior will give banks valuable insights into customer preferences and trends. With this approach, any financial institution can refine its offer and marketing strategies. 
  • Reduced operational costs: By automating the recommendation process, financial institutions can streamline operations and reduce overhead costs. 

How to use recommender systems in banking and financial services? 

As we mentioned, recommender systems can be used across various areas within the banking and financial services sector. This includes:  

  • Personalized product recommendations: Recommenders are great at tailoring product suggestions about investment options or loans based on what individual customers need and prefer. 
  • Customer segmentation and targeted marketing: Thanks to recommender systems and their ability to analyze customer data, banks can create and deliver targeted marketing campaigns. 
  • Portfolio management and investment advice: Providing personalized investment advice and portfolio management services can be much easier with recommenders because they can help better understand your clients’ risk profiles and financial goals. 
  • Customer service improvement: Finally, with recommenders by your side helping them understand customers better, banks can improve customer service through prompt and relevant help and support. 

Challenges and limitations of recommender systems in banking

However, despite their many benefits, recommender systems also pose several challenges for the banking and financial services industry: 

  • Protecting customer data and ensuring compliance with data privacy regulations can pose a challenge with some recommender systems.  It’s worth noting that our system is fully GDPR compliant. Storing relevant consent data in the Solver Data Lake enables you to choose which data can be processed and generate personalized messages for specific recipients. Additionally, you can modify your preferences over time. 
  • Relying solely on past user behavior and historical data may limit the recommender’s system’s ability to adapt to changing preferences and market dynamics. 

To address this second issue, our recommender systems can incorporate or combine event streaming to capture important triggers as signs of customer needs or recognize important live moments, thus merging historical data with (almost) real-time data.  

However, being aware of all the challenges and potential risks for banks, we at Things Solver have developed a new approach that specifically caters to the way they operate.   

Read more about it below.  

The new approach for banking and financial services by Things Solver 

The Things Solver recommender systems make their way into the banking system through the online session’s data. Here’s how that looks in reality:  

An online session is defined as a visitor’s (client’s) visit to the website and its entire activity there – for example: 

  • The path through the website,  
  • Time spent on certain pages, and  
  • The choice of links.  

Based on that data, we try to describe what the client wants – the client’s visit to the webpage for loans or his attempt to make a calculation can send us a signal that there is some interest in taking a loan. 

Once this process is done, banks get a five-product package the client in question has looked at and analyzed. Additionally, the algorithm can recommend another five products the client may find appealing but hasn’t explored yet —highlighting the algorithm’s ability to discern potential interest in these offerings. 

For at least 10% of clients, we can notice the interest even before they show it. That’s not a small number. The percentage would grow if those clients were contacted to keep their interest. 

As you probably anticipate, this is the moment when managers usually become more interested in the recommender systems.  

Yet, these results don’t only make the management happy. Other employees are also happy because, with minimal changes in procedures (major changes are often unwelcome), there’s an improvement in results and efficiency, too.  

What to expect in the future?  

Recommenders in banking and financial industry
Source: Pexels

The AI revolution in the banking/financial services industry has only just begun.  

In fact, several trends will be shaping the future of recommender systems in this industry. We believe that we can expect two major directions:  

  • AI-driven hyper-personalization: In the future (near or far), banks will be relying even more on artificial intelligence to deliver highly personalized recommendations that completely cater to individual preferences and behaviors. 
  • Omnichannel experiences: Also, banks will be integrating recommender systems across multiple channels to provide seamless and consistent experiences both online and offline.  

Recommender systems truly have a great potential for improving the way banks operate within the constantly evolving banking landscape. In addition to helping banking institutions better understand their customers, recommenders can also boost engagement and business growth.  

Are you ready to address your most common challenges and implement the right AI technologies?  

Then you’re in the right place! 

Contact us at or simply book a demo so we can find the best approach for your business.