Chronic low-back pain is an illness that shows a high prevalence worldwide. There is a set of causes that we can attribute to such condition. It does range from muscle contractures, arthritis, radiculopathies, degeneration of the intervertebral disc to pathologies such as cancer. Nevertheless, 90% of such pain has been classified as non-specific. In fact, the treatment for the illness is commonly oriented towards the symptoms. Rarely, it is investigated. To get the most of it, we are proposing a system to identify, analyze and generate a treatment for low-back pain. The system encompasses a set of sensors and algorithms based on ML and AI as the backbone that might allow us to properly identify it and to accordingly generate a diagnosis suggestion. In this trend, an experimental study will be carried out where two groups (observational/experimental) will be managed. The subjects are male and female who are in the age range between 31 to 64 years. For the observational group, the inclusion criteria state healthy people who do not have any low-back pain diagnosis and they are not under any sort of treatment while for the experimental one, whole patients feature low-back pain, whether it has a specific cause or not. By performing such analysis, it is possible to gather enough information to identify, to certain level, the key differences among the several back pain conditions and the possible diagnosis suggestion featuring a large certainty based on the amount of people with the illness measured as possible.