Case study – Piazza CopernicoDesigning Digital Learning Courses with Generative AI
For Piazza Copernico, we developed an LLM-based system that supports instructional designers in creating complex and personalized educational content, while maintaining quality and confidentiality.
Case study – Piazza CopernicoDesigning Digital Learning Courses with Generative AI
Piazza CopernicoDesigning Digital Learning Courses with Generative AI
For Piazza Copernico, we developed an LLM-based system that supports instructional designers in creating complex and personalized educational content, while maintaining quality and confidentiality.
-
Long times for analysis and writing -
Standards and consistency hard to maintain
Creating content for digital learning courses requires complex, consistent, and personalized texts for every learning path. Doing this manually slows down production and makes it difficult to maintain consistent standards across growing volumes, especially when sources are many and fragmented. The need was to speed up without losing control and verifiability.
-
Source-driven generation (RAG)
-
Verification by querying the sources
GRID+ developed a system based on LLMs with Retrieval-Augmented Generation (RAG) methodology that generates complex, personalized texts for every course starting from defined source materials. The model produces content while also allowing users to query the sources to verify and validate the output, making the workflow more controllable and reliable.
-
Concrete support for analysis and writing
-
More focus on instructional strategy and quality
The solution is operational and currently in testing to measure its impact on the workflow. It supports trainers in the analysis and drafting phases, improving human-machine interaction and reducing repetitive tasks. This frees up time to focus on instructional choices, the structure of learning paths, and the overall quality of the content.
What changed
Before and after introducing
Generative AI
Before
Before
Manual analysis and macro-design, with long times.
Hard to speed up without reducing quality and consistency.
Source verification more scattered during drafting.
After
After
AI support in the analysis and text drafting phases.
Source querying to verify the output.
More time for instructional designers to focus on instructional strategies.
