Zenya 7.8.0 Update 3

Every 14 days, we release a minor update for Zenya SaaS. In these updates, we make minor improvements, improve Zenya's performance and fix bugs. Minor updates have minimal impact for end users of Zenya. Changes with greater impact are always rolled out in major updates, which come out about 3x a year and are always announced in advance.

Zenya AI Chat is available 

With this release, we are rolling out Zenya AI Chat to production for customers with a Zenya AI Chat licence. Zenya AI Chat is an AI chat module that is integrated directly into the Zenya environment. Staff ask questions in everyday language and receive answers based on their organisation’s own documents: protocols, policy manuals, guidelines and procedures. 

The core features of Zenya AI Chat are Assistants: configurable AI specialists that administrators can set up themselves for each target group or department. A healthcare organisation can thus set up a separate Assistant for nursing, HR and the works council, each linked to the relevant document collection. 

This release marks the launch of the first version of Zenya AI Chat: from the conversation interface and the configuration screen to the rights management system and the quality of the AI. 

Want to know more? Talk to us about the possibilities

Is your organisation not currently using Zenya AI Chat, but are you curious to find out more about our solution? Please contact your account manager or consultant for further information.

Please contact us

Conducting interviews with assistants 

Users can ask an Assistant questions in everyday language. The Assistant searches the linked document collection, formulates a response and refers to the source documents. If the answer is not found in the sources, the Assistant explicitly states this; it does not fall back on general AI knowledge. This is a deliberate choice to ensure reliability and information governance. 

Zenya AI Chat can be accessed via a browser on desktop, tablet and mobile devices. It has been designed to work effectively on different screen sizes. 

Users can switch between Assistants during a call via the drop-down in the chat, without losing the context of the conversation. This makes it possible to consult several specialists within a single, seamless workflow: a question about a care protocol to the Nursing Assistant, followed immediately by a related HR query to the HR Assistant, without having to start again. This sets Zenya AI Chat basically, separate chat windows for each department. 

If a question can be interpreted in more than one way, the Assistant first asks a clarifying question before formulating an answer. This ensures that answers are better tailored to what the user actually wants to know. 

Administrators can configure up to 20 starter questions per Assistant. Each time a user opens a new conversation, four suggestions are displayed at random. This lowers the barrier to getting started and helps users discover what they can ask. 

Each conversation session is automatically given a short, descriptive title based on the first reply. This makes it easy to recognise the conversation history and back can be found. 

When a user opens a conversation that is older than a certain period, a notification appears stating that the information may be out of date. This helps users to be mindful of the reliability of the answers. 

Setting up and managing assistants 

Administrators set up Assistants via a clear and organised management screen. Without needing to know any specific prompts, they can configure the behaviour, sources and permissions for each Assistant to suit the target audience or department. From role settings and writing style to language level and document collection: everything can be configured via a structured interface. 

There is a separate “Assistants” page available via the Zenya-sidebar, visible to users with administrator rights. Here, administrators can view all Assistants for which they have rights, filter them by name and create new Assistants. 

The configuration page for an Assistant consists of two tabs: 

  • Assistant management: name, description, behaviour, sources and additional instructions 
  • Settings: permissions, copying and deleting 

Instead of a free-text field for a prompt, the configuration page offers a structured interface with four parameters: 

  • Role / Persona: the area of expertise or role that the Assistant takes on (e.g. HR adviser, nursing coach) 
  • Answer length: brief, standard or detailed 
  • Writing style: personal, neutral or businesslike 
  • Language level: from very basic (A2) to expert (C2) 

Administrators can link a document collection and use a free-text field for additional instructions for advanced configuration. 

Rights operate on two levels: 

At system level: 

  • Super administrator: full access to all Assistants and settings 
  • Creator: can create new Assistants 

At Assistant level: 

  • Administrator: can edit the Assistant and assign permissions 
  • Content Manager: can edit the Assistant 
  • User: can use the Assistant (chat) 

When an Assistant is created, the creator is automatically assigned the role of Administrator. This allows for department-specific ownership: only the HR team manages the HR Assistant, and only the quality team manages the Quality Assistant. 

The creation, editing and deletion of an Assistant are recorded in the activity log, including the username, the time and which fields have been changed.  

AI quality and reliability

Zenya AI Chat is built on the principle that answers are always traceable. Assistants respond solely on the basis of their own organisation’s documents and explicitly state when information is not available. Technical measures such as hybrid search and automatic further attempts to locate ensure that relevant sources are found, even when dealing with complex questions. 

Assistants only provide answers based on the indexed sources. If information is not available, the Assistant will explicitly state this. There is no fallback to the AI model’s general training data. 

The search function combines semantic search (based on meaning) with keyword search. This means that users can either ask a descriptive question or search for a specific word or protocol number, and will receive relevant results in both cases. 

If the first search does not return any useful documents, a second, refined search is carried out automatically. This increases the likelihood that relevant passages will still be found, without the user having to do anything. 

For customers who use Vilans In protocols, this document type is available as an AI knowledge source. When Vilans If protocols are enabled for a tenant, they are automatically included in searches. 

Staff can now ask questionsand about processes to Zenya AI Chat and AI Search. The Assistant understands how a proces is structured and formulates answers based on the process structure, not just plain text. This is a significant improvement on the previous indexing of Visio processes, where only plain text was available to AI Search. The new indexing takes the entire process structure into account, resulting in more accurate and context-aware answers. 

New features and improvements
In this update, you will find useful new features and several bug fixes that ensure a smoother experience. Check out what's new here.

RISK

Control Tests that have not yet been reviewed have been added to the “Test review status dashboard”
The Test Review Status dashboard now also shows which Control Tests have not yet been reviewed. Previously, this dashboard only displayed completed reviews.

  • Tests that still need to be reviewed now appear with a grey “Not yet performed” status, just like the “Not yet tested” status for Control Tests.
  • These outstanding reviews are shown in the “Review trends over time” chart and in the “Control details reviews” table, dated according to the underlying Control Test.
  • The “Test review overview” dashboard remains unchanged and shows only completed reviews.

General

General improvements to security and performance
A number of improvements have been made in terms of performance and security. Overall stability has been enhanced. At the same time, security measures have been tightened to better protect data.

Particular attention has been paid to users of “public access”: we have added further optimisations and security checks to make the experience safer and more reliable.

Bug fixes

We have fixed several bugs to improve stability and performance. Want to see all the details?

Click here

When available?

See status.infoland.nl for the exact date when this update will be rolled out to your organisation's Zenya environment(s).

Post published on 19 June 2026 in Zenya SaaS