AI-powered profile analysis according to Griesshaber at Berufswahlschule Uster: from handwritten texts to an Excel evaluation
BWS Uster · Project overview ·
The project at a glance
Evoya AI has developed a web application for the Berufswahlschule Uster (BWS) that automates the profile analysis according to Griesshaber for German. The application digitises handwritten student texts, divides them into minimal sentence-worthy units (MSE, minimale satzwertige Einheiten), analyses them using the Griesshaber method and exports the result to an Excel file.
Teachers save 30 to 60 minutes per profile analysis. They can adapt the AI instructions themselves and test them via a chat agent. The project took two weeks.
- Client
- Berufswahlschule Uster (BWS)
- Industry
- Education
- Use case
- Profile analysis according to Griesshaber for handwritten German texts
- Solution
- Web application with AI handwriting recognition, MSE division, analysis and Excel export
- Project duration
- 2 weeks
- Time saved
- 30–60 minutes per profile analysis
The process in pictures
Initial situation: analysing handwriting by hand takes time
At BWS Uster, the profile analysis according to Griesshaber was a time-consuming manual process. Teachers had to transfer handwritten texts to Excel, divide them into minimal sentence-worthy units (MSE) and categorise them. This was lengthy and error-prone.
Classic OCR technology was of limited help: different handwriting and varying language proficiency among students made text recognition difficult. In addition, interpreting the texts required a great deal of sensitivity from the teachers.
Goal: automate the whole process in one application
The web application was to cover the entire profile analysis process that had previously been done by hand:
- Capture handwriting: no more typing up texts.
- Divide texts: form MSE automatically.
- Analyse: categorise the units using the Griesshaber method.
- Deliver results: output the results directly as an Excel file.
At the same time, teachers were to be able to adapt and test the AI processing themselves.
The solution: from handwriting to Excel evaluation in five steps
The web application combines several AI-supported steps into one continuous process.
1. Handwriting digitisation: AI models recognise and interpret the handwritten student texts, including texts that span several pages.
2. Division into minimal sentence-worthy units (MSE): the application automatically segments the digitised text into MSE.
3. Analysis creation: the units are categorised and analysed according to the Griesshaber method.
4. Output to an Excel file: the results are exported automatically to an Excel file.
5. Continuous optimisation: the AI instructions are managed centrally and can easily be adapted by teachers. A chat agent lets them use and test the application directly.
Results: 30 to 60 minutes less effort per profile analysis
The project was implemented in two weeks. It gives BWS Uster the following benefits:
- Time savings: 30 to 60 minutes per profile analysis.
- Consistency: the AI analysis produces more consistent results.
- Scalability: a large number of students can be analysed without additional effort.
- Flexibility: teachers adapt the AI instructions themselves.
- A basis for more: the analysis forms the basis for development plans, exercises and interactive tutors.
Outlook: development plans, exercises and an interactive tutor
The process does not end with the analysis. The planned next steps are:
- individual development plans,
- personalised exercises,
- an interactive tutor that accompanies students based on their development plan.
Related topics: Web applications · AI process automation
Frequently asked questions about the project
The web application digitises handwritten texts, divides them into minimal sentence-worthy units (MSE), analyses them using the Griesshaber method and exports the results to an Excel file.
Different handwriting and varying language proficiency made classic OCR technology difficult to use. In the application, AI models recognise and interpret the handwritten texts.
Teachers save 30 to 60 minutes per profile analysis.
Yes. The AI instructions are managed centrally and can easily be adapted by teachers. A chat agent lets them use and test the application directly.
The project took two weeks.
Automate manual evaluations with AI
Talk to us about which of your processes could be automated with an AI web application.