Zivver

Meierijstad Municipality

Zivver helps email users determine when an email contains sensitive information and needs to be sent securely. Zivver secures emails using an algorithm.

Last change on 24th of June 2026, at 7:02 (CET) | Publication Standard 1.0

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

General information

Theme

Organisation and business operations

Begin date

2020-09

Contact information

[e-mail]

Responsible use

Goal and impact

The aim of this algorithm is to help ensure that information sent by email is secured appropriately. In other words: not with too few security measures, but not with too many either. The algorithm achieves this by alerting email senders to sensitive information and encouraging them to send this information securely. If they choose to do so, the email is secured using encryption, two-factor authentication for the recipient, the ability to revoke access, and logging.

This ensures that sensitive information is protected, whilst non-sensitive information remains accessible without causing any additional friction for the recipient. The algorithm only affects the process of sending emails and does not impact the substantive legal position of citizens or businesses.

Considerations

The use of this algorithm helps to ensure that emails are sent with the appropriate level of security. Classifying emails using an algorithm is more effective than classification based on a word list compiled through human input or classification by the sender, whilst at the same time the algorithm does not affect the content of an email.

Human intervention

In principle, the user decides for themselves whether or not to send an email securely, based on the algorithm’s recommendation. It is also possible to automate the decision on whether or not to send an email securely, based on the algorithm. In that case, a user can still choose to send an email securely explicitly.

Risk management

The algorithm’s overall performance is monitored by the supplier. If it transpires that the algorithm is producing incorrect classifications more frequently, this is detected by the monitoring system, enabling adjustments to be made to the algorithm. For the organisation using the system, the accuracy of Smart Classification is also visible in an administrator dashboard.

Operations

Data

The algorithm is based on anonymised historical messages, combined with information on whether or not these messages were sent securely. The anonymisation is carried out in several stages, with only aggregated data covering many thousands of messages being stored. In addition, personal data is automatically removed from the data. This ensures that the data is anonymised and that it is impossible to extract the original messages from it.

Technical design

When composing a new email, the system assesses how similar it is to previously sent (secure or normal) emails, based on the terms used in the message and the attachments. If the email is sufficiently similar to messages that are normally sent securely, it is flagged as potentially sensitive. The terminology used in the email is analysed to determine whether it relates to a specific category of sensitive information, such as medical or legal information. Once the email has been classified as sensitive and concerns a subject that the organisation has specified must be sent securely, the user receives a recommendation to send the email securely. 

External provider

Zivver