Streamlining Workforce Management with AI and Visual Analytics

The Challenge

JBM Group is a conglomerate that manufactures key auto systems, electric vehicles and buses. The Group currently has 40 manufacturing plants and four engineering and design centres across 18 locations.

The key challenge confronting JBM Group was the inefficient and time-consuming manual processes for tracking employee attendance and workforce planning across its manufacturing plants. To overcome these bottlenecks, the Group CIO, Ranganathan Iyer, implemented two cutting-edge solutions – an AI face recognition attendance system and a visual analytics solution. Both solutions were developed by Iyer in-house through a subsidiary (Third Eye AI, a JBM Industry 4.0 solution provider) through a team of professionals.

AI-Powered Facial Recognition Attendance System

Prior to implementing the facial recognition system, JBM Group relied on manual attendance marking, which was a slow and error-prone process. This made it difficult to have real-time visibility into employee presence.

The new cloud-based solution, developed and deployed during the COVID pandemic, eliminated the need for manual attendance marking, providing a fast and contactless way to track employee presence. It enabled the company to capture each employee’s attendance and register the time in less than four seconds. This data was then integrated with the company’s SAP ERP system. During development, this solution was deployed seamlessly without any development team’s physical presence across the JBM plants.

“When the 7 a.m. shift starts, we get to know the availability of all the critical resources. It helps us in employee-wise, skill-wise, and machine-wise mapping. Earlier, all this was done in Excel, which took up a lot of time. Now, the data is captured within the software solution, linked to our SAP solution, and consumed by the production department for planning,” says Iyer.

As a result, JBM Group has achieved workforce planning and optimisation while improving productivity and decision-making.

“What used to take 15 minutes for every shift at every plant, now takes only two to four minutes. With the ability to quickly determine the availability of critical machine operators, the production department now makes informed decisions to optimise the workforce,” Iyer says.

“The added advantage of this solution is that we are able to budget the manpower both in number and financial terms. This solution helps business heads to understand the present status of headcount overrun along with cost overrun, which enables them to take real-time corrective action as the visibility of both the parameters is available by the click of a button,” he says.

IBRS believes AI-powered facial recognition technology can be used for employee attendance tracking in manufacturing environments. However, it is crucial to carefully evaluate such systems’ accuracy and potential biases, especially when dealing with diverse workforces, as there could be a risk of false readings for minority groups. In addition, any use of facial recognition technology must be implemented in consultation with staff, with clear limitations on how the technology will be used. In some organisations, the use of AI (be it face recognition or other behavioural mechanism) for tracking staff activities will not be viable due to the culture of the workforce. 

IBRS also notes that the use of AI for staff attendance and activity monitoring needs to take into account the culture outside of the organisation. While such tracking at facilities in many India and ASEAN countries may be acceptable, it would be frowned upon in Australia.  

However, IBRS also expects far higher public tolerance for AI tracking within educational institutions, aged and healthcare facilities and other specialised locations, with safety and convenience outweighing distrust.

Even with these caveats, IBRS notes that there are significant planning, productivity and safety gains to be had by leveraging AI vision. IBRS recommends exploring solutions from a wide range of technologies like radio frequency identification (RFID), movement tracking, and  AI. This multi-layered approach can enhance security, track personnel movement to assist with workforce and workplace optimisation, and ensure compliance with safety protocols.

Visual Analytics

When it came to manpower monitoring, the manual processes previously in place at JBM group did not provide the necessary insights to optimise manpower allocation and address operational inefficiencies. This hindered the company’s ability to improve process efficiency.

“Manpower monitoring is an essential aspect of maintaining quality. Only when the right process and resource are identified, a quality product is manufactured in optimum time. Having too many operators on a machine could lead to pilferage and cost overruns, while too few could result in quality issues and high rejection rates,” says Iyer.

Without a robust monitoring system, it was difficult for JBM Group to track how much time each operator spent at their assigned machine. This made it challenging to identify the root cause of any issues, whether it was an operator-related problem or a machine malfunction.

“To address the challenge of effective manpower monitoring and optimisation, we implemented a visual analytics solution with strategically placed cameras across our manufacturing operations. This solution provided a comprehensive and data-driven approach to managing operator presence and activity,” says Iyer.

The visual analytics solution was designed to alert line managers or plant in-charges whenever there were anomalies in manpower allocation. For example, if there were more or fewer operators than the stipulated number assigned to a machine, the system would trigger an alert.

The solution also tracked the time each operator spent away from their machine, providing insights into the reasons behind their absence. This helped identify whether the issue was related to operator training or the machine itself.

“By leveraging the data collected by the visual analytics solution, we gained valuable insights into process efficiency. The company could analyse operator performance, identify bottlenecks, and make data-driven decisions to optimise manpower allocation and improve overall productivity,” adds Iyer.IBRS notes that integrating video analytics with the Internet of Things (IoT) and programmable logic controllers (PLCs) can provide more comprehensive data for planning plant equipment maintenance as well as identifying workforce training needs. This approach can address known issues and reveal unexpected and invisible problems that humans may miss. Integrating video analytics with other systems like enterprise resource planning (ERP) and manufacturing execution systems (MES) can provide a holistic view of the manufacturing process, enabling better decision-making and optimisation.