We turn your ideas into math and software
We turn your ideas into math and software

Case Study: Intelligent Processing of Real Estate Documents for Finnish Bank

We reduced manual typing required to serve mortgage and real estate customers
up to 80%. That means more time for the customer encounter, less for typing.

Automatically converting scanned documents into structured data was the Bank’s automation dream. Now it’s reality and it all happens safely in their own private cloud thanks to Digisalix small data AI -solution. No more wasted hours on mindless typing by highly paid professionals.

Check out the Case Study to learn how it’s done and if Digisalix is a solution for your automation dreams.


  • 80% less typing means more time to serve the customers.
  • 12 minutes time savings per customer/certificate.
  • Document data is now instantly structured into a database-friendly form.

Even complex data types were extracted, e.g. lists of building renovations.

The small data approach

Instead of using big data, we gave structural hints for the computer to understand the content. Our customer, The Bank, only had to provide us a modest amount of example documents and an access to their private cloud for us to install a custom smart scanning system for “building manager’s certificates” in 8 short weeks. We trained the Digisalix machine learning system with only 848 documents. This small data approach enables efficient setup without compromising the accuracy of the system.

Download the Case Study and learn all about our IDP of real estate documents

We’ll walk you through the whole process in a way that is suitable for both techies and, well, other people too.

Share this article


More posts by Janne Lehtinen

We found our summer trainee

The trainee position for summer 2024 is filled. See you next year? “If you’re into mathematics, algorithms and programming, we might have a perfect internship for you. We expect you to be familiar with

Read More

We teach computers how to read better

In order to process information, computers need it in a structured form. All pieces of data must have strict meaning and format. On the other hand, humans can infer meaning from context and circumstances.

Read More