سیناپس نگار

آخرین اخبار هوش مصنوعی، تحلیل‌ها و ابزارهای AI

Pré-restitution : des outils et de l’intelligence artificielle

Flottes Automobiles

Source published: Site published:

Original article · فارسی

Pre-restitution: tools and artificial intelligence

For fleet managers, ensuring the vehicles are in good condition before they are returned helps to limit any unpleasant surprises. To do this, tools are multiplying, including applications that increasingly rely on artificial intelligence. With results but also limitations that leave plenty of room for human validation.
by Julie Vénier -
Magazine No. 321
127
management tools

PTo avoid discovering the defects of a vehicle during its return, that is, after the last moment, it is necessary to organize by conducting multiple inspections of vehicles upstream within the company. This is the case at Xylem, a specialist in water solutions and technologies, which operates 176 vehicles in LLD and relies on WeProov, a specialist in digital vehicle inspection.

«The technician or collaborator receives a link on their smartphone. Then, the application guides them to photograph the vehicle: such area, such angle, such light, etc. The entire vehicle is thus documented,» explains Ghislain...

PTo avoid discovering the defects of a vehicle during its return, that is, after the last moment, it is necessary to organize by conducting multiple inspections of vehicles upstream within the company. This is the case at Xylem, a specialist in water solutions and technologies, which operates 176 vehicles in LLD and relies on WeProov, a specialist in digital vehicle inspection.

«The technician or collaborator receives a link on their smartphone. Then, the app guides them to photograph the vehicle: such area, such angle, such light, etc. The entire vehicle is thus documented,» explains Ghislain Quinet, purchasing director for the Southwest of Europe at Xylem. Based on these photos, the tool identifies the damage and assesses its cost using the network of repairers offered by WeProov.

Facilitate inspections

« Then, with our Fleeter, we make the final decision to repair, not to repair, or to file a claim with the insurer », continues Ghislain Quinet. Who adds: «WeProov clearly uses AI to identify damage, but the arbitration remains human.».

For its 876 vehicles, David Bally, head of general services and HSE at the industrial engineering group Fives, also relies on a tool: «For almost two years, we have been using the fleet management tool from GAC Technology to launch vehicle inspection campaigns. And this feature is very interesting in the perspective of a return, especially in sites where the fleet manager is not physically present. »The driver is then put to work. «He can now take the vehicle around himself with the app, take the required photographs (front, sides and other areas) and record the mileage or report any anomalies,» anticipates David Bally.

On the other hand, this official remains cautious about AI: «We have not yet actually tested AI as part of our control campaigns; therefore, I cannot say how far it can be used. I do see a potential for helping with decision-making. But at this stage, this has not yet been sufficiently implemented at Fives to yield concrete results. Today, the assessment remains primarily that of the fleet manager concerned and, during the feedback process, that of the controller,», Temper David Bally.

Artificial intelligence...

At Macadam, As a specialist in inspections during restitution, IA is in fact confined to a role of assistance. «Macadam Group has a department that specifically works on AI advancements. We produce more than 1 million photos per month, and this database is used, among other things, to train our automatic damage detection tools. At this stage, IA is an assistance to the inspector. It can signal to him that it thinks it can identify a damage, but the inspector always validates it. We do not let it operate autonomously to determine restoration costs,» explains Vincent Muyllaert, CEO of Macadam.

... to guide users

Moreover, the pre-inspection can also be entrusted to the driver. «The fleet manager can ask the driver to perform a preliminary inspection based on a few photos. With the help of AI, the app can then identify an area that may correspond to damage and guide the user in assessing what might or might not be invoiced. This is very useful for fleets with multiple renters: a 5 cm scratch can be accepted by one and invoiced by another,» recalls Christophe Villain, sales director at Macadam.

The operation is not very different at Mon Petit Carrossier. «When the vehicle reaches the end of its life with our client, we ask the employee, approximately six to eight weeks before the return, to take a series of photographs using our mobile app. We are indeed looking to recover as many views of the vehicle as possible. The images then go into a tool that incorporates an AI component capable of detecting minor damage that is sometimes difficult to see in the photos, such as some bumps,» details Steven Grant, Director of Development.

AI to calculate repairs

AI is also being used in the calculation of repairs, as in the remarketing specialist Dentmaster, which uses AI to generate quotes. «And for fleet managers who have doubts about the merits of a repair, we also provide free quotes on our platform, simply by relying on a few photos. The driver or fleet manager can go directly to carmeleon.fr, the website of our subsidiary Carmeleon, using their phone. From there, they are guided step by step to take photos of the vehicle’s damage so that we can generate the quote. All this first step is completely done remotely,» says Pierre Heyraud, CEO of Dentmaster.

Among our witnesses, CESMAT does not currently use any specific application or AI tool for the pre-inspection of its 250 vehicles. «That may come in the future, but it’s not our immediate priority. We are currently equipping ourselves with fleet management software. Until now, I had a lot of scattered Excel spreadsheets. The goal remains to centralize information, to have more advanced statistics, a dashboard, and alert systems, especially for revisions and various deadlines,» says Muriel Collu-Visconti, T&E Manager at this specialist in nuclear and radiology. The integration of the tool is expected to take place soon. Once this first step is stabilized, we will likely be able to work more on the restitution during the 2027 year.