Motivation

althiqa develops a Saas service for companies to evaluate their artificial intelligence (AI) models in a simple and efficient way.

The adoption of AI is accelerating in the enterprises: the predictive power of machine learning algorithms has created a lot of value and has transformed many industries. In this context, the evaluation of AI models has become a major strategic issue. It is about ensuring that the algorithms which are developed meet the business objectives, the criteria of which are multiple and use cases dependent. Performance metrics (ROI and proxies), the study of discriminatory biases, the transparency of decisions, or even CSR metrics (CO2 consumption, for example), are all key indicators that stakeholders must inspect. In this context, effective communication between data scientists, managers and business experts is essential.

Because the stakes are high, we think there is a crucial need for tooling that allows data scientists to communicate easily and efficiently about the models they develop.

With its SaaS service, althiqa’s mission is to meet this need. We develop a very intuitive tool allowing data scientists to communicate on the key elements of their AI models. The main interface summarizes the different metrics they want to inspect in a form that is accessible to all stakeholders. This allows managers to have real-time visibility on the models under development. It also greatly improves the confidence level of business experts in the algorithms developed by data scientists. In short, our tool ensures that the models developed are aligned with the requirements of each stakeholder.