Getting started with data-driven claims management

Thank you to everyone who joined us and helped us make this webinar a big success. We had close to 100 people signing up from Europe’s leading insurers.

We enjoyed hosting it, and we hope you found it interesting and got some actionable takeaways.

Key take aways

  • Not all data is created equal, some data points can tell you a lot, others very little
  • Technology is an enabler for tapping into unstructured data
  • Technology fuels automation
  • Start taking stock of your data quality
  • Remember the ethics and regulations

Watch the webinar

Use cases

Don’t have time to see the whole webinar? Jump to three real use cases on how data can improve your claims management.

Let’s continue the conversation

We have a proven method for implementing a data-driven claims management methodology within personal injury. Let us analyze your gains. Fill in the form, and we will schedule a digital meeting.

Marko Ahokas

Marko Ahokas
VP Sales
LinkedIn Profile

Johan Larsson
LinkedIn Profile


We got many great questions from the audience. See the questions and answers here.

Are PDF:s structured or unstructured data?

Yes, and no, it depends on how they were created. The PDF format can contain structured data. But often are PDFs:s used to hold for example, an image. That is often the case in medical journals that have been scanned.

How difficult is it to get access and approval to work with customers’ data?

It can be quite hard; you need strict control over what data you need, what you share, and how you use it. But I urge everyone who builds a system to think about the sharability of your data. That will make it easier to share and work with your data.

What kind of key competence is needed to work with data?

It is always a challenge to find the right competence. You need a tech team that understands data and can see patterns. But also data experts like; Data Scientists, Data Analysts, Machine Learning Experts, and AI Experts.

How do you go about streamlining and accuracy of the data?

You need domain knowledge of the data. What it is and where it comes from. You need to know what it is and how to read it. You also need to build testability in your development.
To streamline it, avoid free text fields, clean up the data before capturing it, and make it a natural part of the UX.

Is your nordic database of claims relevant international?

Yes – for example, ICD-codes are an international standard, and an injured knee is always an injured knee no matter where you are.

What programing language is preferred?

Choose the one where you can find good people. Within data science, some languages are preferred. But don’t let the language stop you; it all comes down to the people.

What is the most optimal way to store and treat unstructured data?

It depends on the format. Store it in an original format; PDFs:s should be stored as PDFs:s. It is what you do with that data that matters in the end. You can store them in your standard cloud storage, but take security and compliance into consideration on where you keep them. And when you abstract the data, you can store them anywhere, like a relationship database or a object database.