Workwear as a dataset: what RFID, service, and AI bring
The question of how much enterprises spend specifically on protecting people is currently being actively discussed in the industry. In September, ASIZ analyzed annual and non-financial reports of more than 100 Russian companies. It turned out that the share of PPE in the occupational safety budget varies from 8% to 88% among different companies, partly due to different accounting perimeters: for example, at Mosenergo and TGK-1, laundry costs are included in occupational safety expenses. The Association proposes making the disclosure of PPE expenses a separate line item mandatory for large companies so that safety can be discussed using comparable data.
The service model helps obtain such data at the individual enterprise level: accounting is maintained not only for purchases but also for the entire life cycle of an item.
“When we see the entire life cycle of a product, the very logic of managing workwear changes. We can move from counting the number of purchased items to understanding how much it actually costs to reliably provide one person with functional clothing,” says Burmistrova.
The approach is based on RFID. Each item receives an individual tag, and the system records its movement: issuance, wearing, return, washing, repair. For the employee, everything boils down to three actions: apply the pass, take a clean set from the smart cabinet, and return the used one. Meanwhile, the system knows who received the item and when, how long it has been in use, and where it is currently located.
At “BTK Sfera,” digital accounting is integrated into a service model that combines clothing issuance, industrial washing, repair, replacement, and disposal. Locers and vending cabinets can be installed at enterprises for this purpose.
But the tag is just a way to collect data. Value appears when thousands of usage histories accumulate. They show how long different models actually last, which sizes remain unused, at which sites clothing wears out faster, and how many sets need to be kept in circulation.
The tangible effect is demonstrated by Russian practice, even without RFID and smart cabinets. At Gazprom Drilling and Rusgazdrilling groups of companies, a system for accounting, purchasing, and issuing PPE based on “1C: Occupational Safety CORP” was implemented. It covered more than 26,000 employees, and over 15,000 nomenclature positions were consolidated into a single database. According to published project results, material inventory volume decreased by 18%, material resource costs fell by 15%, and warehouse inventory turnover increased by 28%. If digitization of accounting alone yields such results, then data on the actual usage of each item opens up even more opportunities.
The next step is forecasting. The system can account for the number and professions of workers, their sizes, seasonality, clothing turnover, and the speed of its return from washing and repair. Then, instead of a stockpile “just in case,” the enterprise calculates the quantity of items based on actual consumption.
The question asked by the purchaser also changes. Instead of “how much does a set cost?”, they ask: “how much does it cost to provide one employee?” This metric includes service life, number of sets per person, maintenance, repair, logistics, and reserve stock. For an occupational safety specialist, this is also a convenient figure: it can be compared between sites and year over year.
For BTK Group, such data is especially valuable because production and service are in the same hands. Information flows not only from the factory to the worker but also back.
Suppose that for workers of one profession, the same zone of the garment tears most often. For the service department, this is a routine repair. For developers, it is a signal to check the design, material, or add reinforcement.
On large datasets, one can compare how the same product behaves in different industries, what happens to the material after dozens of wash cycles, and under what conditions wear occurs faster.
“When production and service are connected, usage statistics can return to developers. We begin to receive knowledge not only from laboratory tests but also from real-world product usage,” notes Burmistrova.
As a result, instead of an average service life, there is a precise answer: how long a specific model lasts for a worker of a specific profession in specific conditions.
RFID shows the item’s history. The next level is computer vision, which can assess its actual condition. This is still an area of research, but initial results are already available.
In one scientific study published in 2026, scientists collected a dataset of 1,974 images of medical textiles with blood, medicine, and urine stains, as well as clean samples. For blood and medicines, recognition accuracy was about 88–90%. Dried urine is almost invisible to a regular camera, so it was pre-treated with a chemical reagent, increasing accuracy from 3% to 86%. The practical task of the study was sorting medical textiles before washing.
Another study used hyperspectral imaging. Researchers contaminated cotton, polyester, and blended fabric with technical carbon, protein, and oil, obtaining 120 samples with 1,200 analysis zones. The best model identified the type of contamination with 99.58% accuracy. It should be noted that this work focuses on sorting textile waste for recycling, and the result was obtained under laboratory conditions.
Nevertheless, the direction is clear. If visual analysis is combined with the item’s digital history — the owner’s profession, site, material, past service cycles — the system will be able to decide on its own whether to send the clothing for regular washing, additional treatment, repair, or write-off.
Such data is not only about logistics but also information about working conditions. If clothing from one site regularly returns with characteristic contamination or more often requires special treatment, this is a reason to look closely at the site itself.
A similar conclusion follows from foreign studies of firefighters’ combat gear. Contamination with hazardous substances increased after every call-out and was highest among those who worked directly on extinguishing and search-and-rescue operations. Washing with soap, water, and a brush significantly reduced some contaminants but had almost no effect on the release of others, volatile ones. Hence the researchers’ practical recommendation: determine the order of washing based on the task the person performed during the call-out. Essentially, this is clothing management based on data about human work.
The worker themselves does not need to understand how RFID, forecasts, and analytics are arranged. Their scenario remains simple: hand in dirty clothes and receive clean and functional ones.
“Ultimately, our client is a person who works in our clothing every day. Therefore, technology must be complex inside but maximally simple for the employee,” says Burmistrova.
The smart cabinet is only the visible part of the system. Behind it lie accounting, maintenance, inventory forecasting, wear data, and, in perspective, automatic assessment of product condition. Under this approach, workwear ceases to be merely a consumable: each cycle adds data on service life, loads, and contamination. And the key figure becomes the cost of reliably providing one person with safe clothing.
“BTK Sfera” is a division of BTK Group focused on comprehensive provision of industrial and corporate clients with work, corporate, and protective clothing. The company accompanies equipment throughout its entire service life: analyzes working conditions, develops and supplies clothing, issues it to employees, provides professional care, repair, replacement, and accounting. BTK takes on accounting, storage, maintenance, and clothing turnover, so the client does not need to hold excessive warehouse stocks and independently manage the size range. Corporate clothing can also denote personnel divisions and functions: BTK implemented this approach in a project for “Baltika.”
Other news
The XI All-Russian Occupational Safety Week has begun its work at the Federal Territory "Sirius"
By 2029, the labor shortage in the industry could reach 1.6 million people
"BTK Group" to Prepare a Reconstruction Project for Production Buildings in Prokhladny
Workwear as a dataset: what RFID, service, and AI bring
"Life Form": Premiere of a Film About Russia's First Industrial Workwear Show Held at the "Khudozhestvenny" Cinema