Foris Intelligence

Combining classical techniques with agentic AI that understands real educational contexts, academic complexity is translated into a flexible and adaptable platform.

Intelligence Hero Background

AI designed to scale

Foris understands AI as a vehicle for solving complex problems in creative ways. We decompose academic planning into multiple subproblems addressing them through a hybrid approach that combines metaheuristic techniques, classical algorithms, Machine Learning models, recommendation systems, optimization and interaction with LLM-powered agents. These capabilities, inspired by the literature and adapted to the university context, are uniquely integrated to achieve impossible results for traditional algorithms or sole human intervention.

Natural language to code

Foris has created an intermediate layer that translates institutional rules into a language understandable by algorithms, ensuring flexibility and scalability. In addition, "we integrate agentic workflows to connect user’s natural language with university operational processes, enabling a more direct and intuitive interaction.

Our approach is grounded in continuous learning a strong engineering culture, and a human-in-the-loop model to ensures control and transparency across AI systems.

AI Dimensions

An intelligence framework that integrates advanced analytics, agentic capabilities and optimization models to address academic complexity from multiple angles.

Data Analytics & Machine Learning

A range of Machine Learning and Data Analytics techniques and algorithms are applied to analyze and predict key phenomena across each institution’s student and administrative processes.

LLM & Agents

Agentic workflows and assistants are designed to enable complex queries and task execution through natural language, accelerating technical work and simplifying interaction with the platform.

Planning and Optimization

Optimization algorithms are combined with proprietary heuristics to solve academic planning, scheduling, and university resource allocation problems.

Use Cases

AI as the Core

AI-driven approach to automate key processes, reducing friction and elevating institution's experience.

1. Prediction & Forecast

  • Prediction of student-related phenomena (e.g. dropout)

    Flexible classification models are used within an AutoML framework to identify the most relevant variables and select the model that best fits your data. This enables timely prediction of key phenomena, providing early warnings and clear alerts to support proactive action.

Prediction & Forecast

2. Prediction & Forecast

  • Risk factor analysis

    Dropout risk profiles are built by grouping variables into risk factors and applying models to identify combinations of characteristics that explain increases or decreases in risk. The result is an explainable structure that highlights the conditions driving higher risk, enabling more precise decision making.

Prediction & Forecast

3. Prediction & Forecast

  • Demand forecasting

    The academic demand forecasting engine is based on simulation algorithms that model student enrollment behavior while accounting for institutional rules (prerequisites, credits, constraints, etc.). This approach delivers highly accurate seat forecasts with explainability that is understandable to administrators, even at the individual student level, optimizing class scheduling and enabling data-driven demand planning decisions.

Prediction & Forecast

4. Data Analytics

  • AutoML Platform (Machine Learning)

    The AutoML ecosystem connects natively with data pipelines and enables the training and deployment of Machine Learning models without extensive manual intervention. For universities, this translates into a real-time predictive analytics infrastructure capable of anticipating critical phenomena, such as student dropout or administrative trends, turning raw data into actionable intelligence with reduced implementation time.

Data Analytics

5. Planning & Programming

  • Allocation

    A hybrid approach to scheduling is applied, integrating optimization metaheuristics, search algorithms, and configurable constraint handling through dynamic rules. This engine ensures compliance with critical parameters, such as classroom capacity, faculty workload, and course compatibility, while optimizing infrastructure utilization, schedule quality, and hour allocation.

Planning & Programming

6. Agentic Workflows

  • Through agentic workflows and proprietary tool development, highly complex technical configuration processes are transformed into intuitive natural language interactions or low-code environments. This architecture orchestrates intelligent workflows that automate the definition of institutional rules and the parameterization of algorithms, without requiring specialized technical expertise from the user.

Agentic Workflows

7. Rules Engines

  • Our architecture integrates a proprietary Domain-Specific Language (DSL) connected to a rules execution engine that models business logic in an agile and modular way within core algorithms. Its compatibility with Large Language Models (LLMs) enables complex institutional rules to be created, validated, and adjusted through natural language, allowing universities to modify academic and administrative policies in real time without relying on technical teams or requiring code-level interventions.

Rules Engines

8. Agentic Assistant

  • Student and Administrative Support

    An intelligent assistant supports students and administrative staff throughout the enrollment process by interpreting natural language queries, responding with real-time academic data, and proposing optimal adjustments to course selection, sections, and administrative preferences. Each recommendation is automatically validated by a business rules engine that ensures strict compliance with institutional policies.

    The result is a personalized, self-service experience that reduces operational workload, guarantees academically

Agentic Assistant

FAQs

IA

What we’re always asked about our technology :)

How do your solutions integrate with our existing systems?

Do we need fully clean data to use your AI models?

How explainable are the models you use?

Does the technology require advanced technical expertise from the university?

Does your AI replace processes or teams within the university?

Does the solution work if we have multiple campuses, programs, or academic calendars?

What happens if our academic policies or institutional rules change?

Does your technology scale as the institution grows?