Our powerful machine learning categorisation models are continuously evolving to ensure a more complete, accurate and useful data capture. While similar at the core, every individual’s behaviour and categorisation preferences differ. Through machine learning models, we can tune-in to both population-wide and individual categorisation preferences to consistently generate better quality data.


Always learning

Ever-changing consumer behaviour means that transaction categorisation needs to be smarter, faster and adaptable. Rule-based categorisation is often incomplete or inaccurate, leading to irregular patterns and unreliable results for the end-user. Tink’s machine learning models resolve discrepancies and can fill in the gaps to deliver more complete information. We continuously and safely test, compare and combine smarter models to ensure the best user experience.

Categorisation of transactions.png

How it works

  • Tink’s categorisation model is scalable because it is designed to work with data that comes from multiple sources, even when it includes incomplete or limited metadata, resolving discrepancies and filling in the gaps.

  • The model is always learning, so data is delivered in a way that suits the individual’s preferences. When a majority of similar transactions have been re-categorised that a global change is made, ensuring that users can safely categorise specific one-time expenses without unnecessarily impacting everything else.

  • We use functionality that combines information from different end-users without sharing any sensitive data, while also continually testing and comparing new models to refine the customer experience.

See it in action



Major Nordic bank SEB developed its own PFM for private customers based on the Tink platform. Our simple APIs allow them to integrate all PFM functionalities, such as spending categorisation, as well as push notifications for real time advice and warnings, right inside their regular Mobile Banking app. SEB's app is now the most popular channel for private customers in Sweden, with more than 50 million visits every quarter.  

“Tink's technology is an important building block in SEB's promise to provide state-of-the-art digital banking and a world-class customer experience.”

- Christoffer Malmer, Co-Head Corporate & Private Customer Division.

Categorisation of transactions .png

Need more advice on options for your business? Get in touch and we’ll help you find out how Tink can create a solution tailored just for you. Or continue exploring our other products Account AggregationPayment Initiation and Personal Finance Management.