Alan Ferbach, a graduate of Polytech Nice Sophia Antipolis, is the co-founder and CEO of Videtics, a company established in 2019 with his partners Alexandre Reboul and Pierre-Alexis Le Borgne, specializing in AI-powered video analysis. They have developed solutions designed to leverage video streams to improve security, mobility, and space management in smart cities, private companies, and more recently, in the retail sector.
“We observed that 99% of the streams generated by cameras are never viewed or analyzed. Companies and local authorities possess a wealth of untapped data,” explains Alan. Videtics employs deep learning to identify objects, behaviors, and trends from this dormant data.
In April 2022, Videtics raised €2 million in funding from several investors and Bpifrance. They have 19 employees.
What unique aspect of your sector drove you to use AI?
Alan identifies two major areas that motivated the integration of AI: security and mobility.
Videtics helps its clients detect unusual behaviors, such as intrusion or a vehicle traveling the wrong way. The goal is to reduce the cognitive load on human operators while making video surveillance systems more reliable. “We’re not seeking to replace humans, but rather to enable them to focus on important tasks,” he specifies.
The mobility aspect is equally crucial, particularly for local authorities. “We study how citizens use public spaces to help cities better plan their territory.” For example, Videtics provides data on pedestrian flow in specific areas, enabling infrastructure placement and traffic optimization.
More recently, Videtics has turned toward the retail sector, with promising results. “In stores, our analyses help understand customer behaviors: which aisles they frequent most, where they remain stationary, or which products attract their attention,” explains Alan. This information helps retailers position staff at strategic locations or optimize their advertising campaigns in the most popular areas.
Which AI solutions have you selected?
Videtics has chosen to integrate deep learning technologies, adapted to specific challenges such as object detection and video stream analysis. The team develops its software in-house, which allows for complete customization according to client needs.
One of their key strengths is their ability to reduce false alerts in security systems. Alan illustrates this point with a concrete example. “In a private villa in Ramatuelle, fully equipped with video surveillance and with 2 golden retrievers, our AI helped reduce false alerts from 200 per day to just one or two per week.” This type of innovation ensures reliability and operators less unnecessarily solicited.
How did you overcome cultural and human challenges during AI integration?
Alan emphasizes the need to demystify AI for clients. “Many expect unrealistic performance, like detecting a Swiss Army knife at 200 meters. We must explain what is possible and what belongs to science fiction.”
Videtics supports users so they view AI as a decision-support tool rather than a technology replacing humans. “AI relieves operators of repetitive tasks, but the final choice always remains theirs.” This approach reassures them and ensures progressive and controlled adoption of the solution.
What technical challenges did you encounter when implementing AI?
The Videtics team had to overcome several challenges, particularly regarding efficiency. “Our solutions must be capable of analyzing large quantities of video streams without requiring costly or energy-intensive infrastructure,” explains Alan.
To achieve this, Videtics optimizes its AI models. They reduce their size and complexity while maintaining their performance. “We use techniques to decrease the number of parameters, which limits energy consumption and improves efficiency.”
Another major challenge is selecting and labeling relevant data. “Among hundreds of thousands of images, it is crucial to identify those that will bring real added value to model training,” emphasizes Alan.
What positive changes have you observed in your team’s dynamics thanks to AI?
At Videtics, AI has helped them streamline internal processes, particularly for data management and analysis. “We have a very technological culture, which allows us to focus on innovation while progressively integrating commercial and marketing aspects. I also use generative AI for RFP responses,” explains Alan.
For clients, the benefits are numerous. The data produced by AI serves to make informed decisions and improve the performance of existing systems. In retail, for example, video stream analyses have helped better decipher customer behaviors and optimize human and advertising resources.
What essential advice would you give to SMEs hesitating to take the leap toward AI?
Alan recommends that SMEs start by structuring their data before integrating AI. “Identifying data flows, organizing them, and understanding their potential is a key step.”
He also advocates starting with modest, targeted projects to test the technology. “It is crucial to select employees motivated by innovation and involve them in these initial projects.”
Finally, he emphasizes the importance of asking the right questions to AI tools. “As with a search engine, the quality of results often depends on the precision of the request.”
