PiCO – Intelligent Tutoring System

Specialized AI system for learning support within WBT courses.
PiCO is technically an Intelligent Tutoring System trained for advanced and adaptive teaching support, as it acts as a teaching assistant by varying teaching strategies along the way based on the content being delivered, the related objectives and individual needs.

There are various support strategies:

• At the start: supports the motivation and objectives of the course
• While studying: work on the process of understanding and memorization
• Upon completion of a topic: Supports the participant in self-assessment
• At the end of a learning block: consolidate understanding through practice tests.

By moving beyond the logic of responding to course content, PiCO promotes adaptive learning, as:

• integrates knowledge of the participant's needs into its actions,
• enhances the study path undertaken,
• adapts its action to the individual.

PiCO therefore acts to reinforce the learning of the individual student in a personalized way.

PiCO also won an award Comenius-EduMedia-Siegel, awarded in the category “Didactic Digital Media (DDM)” (Educational digital media).

Pico Learning Spaces

In the landscape of AI teaching assistants, Pico Learning Spaces proposes a substantial paradigm shift in learning support compared to AI teaching assistants that focus on reformulating training content, adapting language, examples, and supporting users in learning formal content prepared by experts. Pico Learning Spaces introduces a methodological layer studied with the contribution of our scientific partners, to give coherence, relevance, and validity to adaptation. Pico Learning Spaces is an agentic system that implements the latest learning theories to provide students with a truly adaptive experience and organizations with the ability to utilize adaptive learning while maintaining governance over the training system. Pico Learning Spaces offers an agentic environment for adaptive learning that manages educational profiling and targeted study sessions supported by PiCO. Integrated with unified domain representation, adaptive assessment systems, and a multimodal learning bot, Pico Learning Spaces is a learning environment complete with individual and class reporting tools, essential for teachers and organizations to analyze and compare different study paths, with a primary focus on skills.

The resulting learning experience is competence-based and at the same time adaptable: contents, sequences and methods are modulated according to the actual learning needs of the individual participant, while progressive checks ensure the achievement of the objectives at each juncture of the process.

Evaluative multi-agent conversational simulator

Multi-agent gym for practical test simulation. The gym offers an operational scenario to perform specific tasks (e.g. objection management, sales, project management, meeting management, etc.).

The simulation can be played multiple times to experiment with alternative scenarios (e.g. different types of sales). It can be played in assisted or evaluation mode.

At the end of the simulation, a specific report on the progress of the test is proposed, useful for the participant to understand their areas of strength and improvement.

LAIVEcase

Conversation simulations are a very useful tool for including practice and assessment in online training programs. To enable the customization and design of interactive teaching simulations, the LaiveCase tool offers a back office organized into a guided simulation scenario design process, integrating all the components needed to create an effective and structured teaching simulation.

Different types of simulations can be designed, choosing the level of multimodality, configuring the agents and the type of orchestration, as well as the LLM model(s) to use (private vs. API vs. mixed models), and even configuring the gameplay (exercise vs. evaluation). The system integrates agentic tools for defining the evaluation model specialized in the story.

The design process concludes with a preview of the learning environment to directly test the design prompts and refine the design.

SmartRAG – Retrievable Agent

Agent in LMS specialized in providing access to all course knowledge.
Through a prompt-driven search approach, the participant can consult the educational resources associated with the courses in which he is enrolled.

Through the RAG (Retrivial Augmented Generation) model developed in Piazza Copernico the student can:

• Locate information within courses
• Ask explanatory questions based on the information provided by the courses
• Use the information to prepare microlearning objects (summaries, questionnaires, tests, etc.) to consolidate knowledge
• Choosing new courses to take

All the contents of our course catalog can be explored through SmartRAG.

SmartRAG can also be used to manage course design processes, as it provides the ability to upload company documentation to be processed for instructional design purposes.

GIW – Gender inclusive writing

To ensure support tools for the correct and ethical use of language, Piazza Copernico has adopted and made available a gender-inclusive writing tool.
The GIW tool allows you to enter text (course content, emails, speeches, etc.) and analyze sentences that include gender marking using a specialized classifier.
For such sentences, the tool offers two types of inclusive rephrasing to choose from to update the text.

GIW was born from a research collaboration with FBK – Fondazione Bruno Kessler and uses algorithmic classification models trained in Italian. This allows the task of detecting gender in Italian to be avoided solely to LLMs, but rather uses a model specifically trained for the task.

Nick – LMS Helper

Nick is the platform assistant who supports participants in managing typical helpdesk issues.

Trained with typical first level assistance cases, Nick offers immediate support to the participant in understanding the problem encountered, classifying the problem via AI and offering an instant personalized response with solutions and operational instructions.

The agent can also be trained with case studies typical of the customer's platform or with other support cases (e.g. registration rules, certification methods, etc.) specific to the context of use.

LearnalyzeR

It is a tool of Learning Analytics integrated into the platform LMS Labe-l Academy and can be integrated via API with other LMSs. It uses composite statistical indicators to observe overall course performance and individual study behaviors. It integrates predictive analytics and anomaly detection tools to identify potentially critical courses and intelligently support tutoring efforts.

The advantage of learning analytics is to observe courses of different content, duration and methods through a uniform measurement system to provide tutors with operational indications for intervene early on potential risks of non-achieving compliance requirements of the courses and at the same time guide tutors and referents in organizing better learning conditions for individual participants.

The analysis and observability model is configurable in terms of variable weights and observation time window, to better adapt to the analysis needs of the individual organization.

Semantic Case Reporting

Semantic analysis software that allows you to analyze document data and tabular information to identify strategic insights.

Through NLP models integrated with Generative AI, Semanticase Reporting helps to read textual data, reducing the effort and complexity of analysis.
It provides topic analysis, time trends, sentiment analysis, but also language metrics.

It allows each operator to start the analysis through a user-friendly system, inserting the intervention of the subject matter expert, not in the onerous phase of classification and organization of the data, but directly in the most important phase of interpretation of the results.

It can be used for data analysis in the HR world, in particular climate surveys, performance management, talent acquisition and HR reporting.