Artificial intelligence (AI) is playing an increasingly important role in quality management and compliance – including at ManualMaster Trevally. During our annual conference on 18 November 2025 at Theater Flint in Amersfoort, AI appeared no fewer than three times on the programme.
Three colleagues – Hugo, Job and Johan – showed how AI already plays a role in ManualMaster Trevally today and which opportunities this offers for organisations that want to use their quality information more intelligently.
AI chatbot in Trevally: fast from question to paragraph
Job is closely involved in the development of the first version of our integrated AI chatbot, built on Azure AI.
This chatbot runs entirely within the boundaries of the Trevally application and does the following:
- You ask a question in normal, human language
- The chatbot responds based on your prompt
- It shows quotations from documents in Trevally
- You immediately receive the URL to the document
- After clicking, you jump straight to the correct paragraph, which is also highlighted
This way, you can find context in your formal system at high speed, without endlessly searching through folder structures or document versions.
Important:
We do not let the AI interpret. The chatbot presents quotations and refers to the corresponding documents. In the highly regulated environments in which our software is used, automatic interpretation based on a single question is a clear no-go: a misinterpretation can have major consequences.
Export data, combine sources: AI outside Trevally
During the plenary session, Johan showed how, with the help of Cody AI, you can build a chatbot that interacts with a vectorised dataset exported from Trevally.
- The data is exported from ManualMaster Trevally
- Outside the Trevally environment, you can link this dataset to data from other systems
- Ideal for organisation-wide AI projects that go beyond Trevally’s document-driven environment
Although his example used an exported dataset, it is of course also possible to make the results available again in the Trevally dashboard for the integrated chatbot within the application.
This creates a bridge between secure document management within Trevally and broad AI applications across the entire organisation.
Data security and AI: Hugo on security and compliance
Data security is our top priority. In a parallel session, our Data Security Officer Hugo discussed:
- The protection of datasets within Trevally
- The risks and preconditions when deploying AI
- How we comply with laws and regulations and internal security requirements
For organisations in and around the Netherlands and the rest of Europe, this is a reassuring thought: our AI applications are developed with security-by-design as a guiding principle.
“He prompts, he prompts”: the art of a good prompt
Everything starts with the basics: how do you write a good prompt?
In a fully booked parallel session titled “Working Smarter with AI”, Job guided participants step by step through:
- The characteristics of a strong prompt
- The reasoning behind that structure
- How to have an iterative conversation with AI and refine the result
The session did not focus solely on Trevally, but on prompting in general. Along the way, topics such as ethics, transparency and the sustainable use of AI naturally came up.
AI as a fixed part of the Trevally roadmap
The use of AI within our application is now a fixed component of the ManualMaster Trevally roadmap.
During the conference, we asked participants who would like to contribute ideas about future AI functionality. Were you unable to attend, but are you interested in AI in relation to quality management, compliance and risk management? Then join one of the masterclasses of the ManualMaster Trevally Academy.
