iGenius approach to AI is built around people. Always.

By building AI around people and not data, we aim to democratize knowledge and enable decision intelligence for Business and Public Administration.

‍Recent recognitions:
Conversational Analytics company in the Gartner’s 2021 Market Heat Index.
Cool Vendor in AI Core Technologies, as well as listed in Gartner’s 2021 Market Guide

iGenius Expert Models

People have different business needs and data skills. That’s why we developed two AI architectures:
GPT For Numbers, the technology behind Crystal, and GPT For General Knowledge that propels the creation of Custom Large Language Models perfect for business-critical use cases.
CRYSTAL

GPT for Numbers

A unique AI architecture designed for business-critical and regulated industries, to improve their decision making process, 100% hallucination-free. 

This architecture is based on Small & Wide Language Models, trained only on specific tasks related to numbers and analytics. Our AI Brain is made of different models that act as neurons, all connected into an AI-Chain. 

On top, to manage this collection, there is an AI Brain Orchestrator, that implements AI Guardrailing to minimize errors and avoid hallucinations. This guarantees security and reliability of informaton at any time.
CUSTOM LARGE LANGUAGE MODELS

GPT for General Knowledge

This AI Architecture focuses on democratizing knowledge, that encourages a responsible use of public data and information. 

Our GPT for General Knowledge is based on Large Language Models (LLM), that we perfected to minimize biases, guarantee fairness, filter inappropriate data, and respect copyright, as the European AI Act recommends. 

We are developing a Foundational Large Language Model that will be released as open source under the MIT license. 

GPT For Numbers

GPT-N is AI that's purpose-built for analytics.

Custom-trained on your lexicon &
metadata.

Enable conversational analytics for
existing data sources.

Sales Over Time for United States

Show me a forecast of sales over time in our United States sales data.