Quality-oriented testing

Groningen Mining Damage Institute

Quality-oriented testing enables automated testing, based on business rules, of certain advisory reports.

Last change on 23rd of January 2026, at 11:21 (CET) | Publication Standard 1.0

• Publication category: Other algorithms
• Impact assessment: Field not filled in.
• Status: In development

General information

Theme

Law

Begin date

Field not filled in.

Contact information

[e-mail]

Responsible use

Goal and impact

The algorithm was developed to make damage reports milder, more humane and easier to review. Previously, each advisory report was fully checked by an IMG staff member. With quality-based testing, the system categorises the risks of the quality of advisory reports and recording reports. This is done on the basis of rules (business rules). Through this classification, the system determines which reports require minor checking or no human checking at all. This allows for faster and more focused assessment.

Considerations

Checking advisory reports was always done entirely by hand since the IMG's inception. This process was reliable, but it cost team Review a lot of time, especially during busy periods. Manual checking can also involve errors. To make the process ready for the future, Team Review wanted to work in a more modern, efficient and data-driven way.

With data-driven working, Team Review can perform quality checks more systematically and clearly. This is done through checkpoints defined in rules (business rules). This automated support allows reviewers to better focus on the more difficult advisory reports and other special cases. Thus, data-driven testing does not replace human control, but rather makes it better.

The three main benefits are:

• Possible errors are found automatically, improving the quality of advisory reports;
• Difficult reports are immediately recognised, so they can be checked more specifically and with extra attention;
• Low-risk advisory reports can be checked faster, speeding up the process.

Human intervention

In quality-oriented testing, there are four testing options:

1. A full human-based test;
2. A test based on rules (business rules);
3. An internal tests based on rules;
4. No human test.

Article 22 of the AVG deals with situations where a fully automated decision with legal effect is taken. In the first three ways, there is always a human review, so this does not fall under Article 22 of the AVG. In the fourth way, it proceeds slightly differently. In this case, an IMG employee does not check the damage reports and no rules are established beforehand. Again, Article 22 of the AVG does not apply. The final 'decision' is the IMG's decision based on a claim application. Besides checking the advisory report, there are many other factors and steps in the process that affect the final decision.

Risk management

As explained above, the IMG's final decision is controlled by a human in all cases. The algorithm only helps make claims cases milder, more humane and easier to handle.

Legal basis

The Groningen Mining Damage Institute is authorised to deal with damage according to article 2 of the Groningen Temporary Act.

Links to legal bases

Article 2, Groningen temporary law: https://wetten.overheid.nl/jci1.3:c:BWBR0043252&hoofdstuk=2&artikel=2&z=2025-07-17&g=2025-07-17

Operations

Data

Advisory reports: including substantive errors (mis-filling of information by experts/errors in calculations), presumption of evidence.