The production management team member struggling with the handwritten production logbook from the morning is a common sight in most small and medium-sized manufacturing companies that have adopted the Excel inventory management system.
The production logbook is received from the field and entered into Excel one by one, and the current inventory is calculated by subtracting the quantity shipped the previous day. This seemingly simple task usually takes about an hour.
The purchasing team received a request for data for ordering raw materials. However, when they opened the Excel file, it appeared that someone had accidentally changed the formula on the field yesterday. There were also cells showing inventory quantities in negative figures, so they had to check everything from the beginning.
We often receive reports that production was about to be stopped due to a lack of materials on site. Excel showed that there was enough stock.
When we went to the site, the actual stock was different. In the end, we had to urgently bring materials from another factory.
Does this sound familiar to you? Inventory management using Excel is now clearly showing its limitations.
In particular, as global supply chain uncertainties are increasing and the accuracy and speed of inventory management are emerging as key factors in corporate competitiveness, these limitations are becoming a risk that cannot be tolerated any longer.
In this article, we will take a closer look at the problems with Excel inventory management that actually occur in manufacturing sites and specifically address how these problems can be solved by introducing an AI inventory management system.
In particular, we will talk about the fundamental changes and innovations that AI systems can bring, focusing on the various difficulties that practitioners face and practical solutions to solve them.
💡 Recommend Contents Practical limitations of ERP systems, how to overcome them? >>
The production logbook, which is written manually at the manufacturing site, is the starting point for inventory management, but it is also the source of the biggest errors. There are various problems, such as misreading due to different handwriting of each on-site worker, unclear records during nighttime production, and omissions in a rush.
In particular, in manufacturing sites where three shifts are in operation, omissions and errors frequently occur during the handover process, and the possibility of secondary errors increases when entering them into Excel.
A single input error affects the reliability of the entire inventory data, which in turn causes a chain of problems from production planning to material ordering.
To prevent such errors, on-site personnel perform double and triple checks, which takes up a significant portion of their workday just for simple data entry.
The biggest flaw in Excel-based inventory management is the lack of real-time data.
This is why it is difficult to respond immediately to urgent orders from the sales team or inventory inquiries from the production team. Even while checking the inventory quantity, materials are continuously being input on the site, and these changes are not reflected in real time.
In particular, it is even more difficult to accurately identify the inventory of 30-40 materials that go into a single finished product for products with complex BOM structures.
In addition, the restriction that multiple departments cannot access the same Excel file at the same time makes real-time information sharing impossible, which in turn leads to delays in decision-making.
Another serious problem with Excel-based inventory management systems is the difficulty of maintaining data integrity. In the process of sharing and using a single file by multiple people, formulas are often corrupted or reference relationships are broken.
In the case of complex pivot tables or sheets containing reference cells, even a minor mistake can threaten the reliability of the entire data. These problems can lead to incorrect decision-making, beyond simply numerical errors.
For example, if the inventory is shown as higher than it actually is due to a calculation error, the timing of material ordering may be missed, which may lead to production delays. Conversely, if the inventory is shown as lower than it actually is, unnecessary urgent orders may be placed, leading to waste of money.
In addition, due to difficulties in backing up data or managing versions, it is often impossible to restore the previous state when a problem occurs.
The discrepancy between the quantity of inventory shown in Excel and the actual quantity of inventory on site is one of the biggest headaches in the process of managing inventory in Excel for manufacturing companies. This discrepancy occurs for a variety of reasons.
The main causes are that real-time input is not performed when materials are input or defects occur, emergency use of materials at the site is not recorded, and regular inventory audits are difficult. In particular, these discrepancies are more serious in manufacturing sites where multiple processes are carried out simultaneously.
As a result, the reliability of production planning is reduced, and situations arise that require urgent material flexibility or sudden adjustments to production schedules. This is a serious problem that directly leads to a decrease in productivity and an increase in costs.
The creation of monthly inventory status and trend analysis reports reveals another weakness of the Excel-based inventory management system.
The main problems identified are the slow processing speed due to the system load that occurs in the process of handling vast amounts of data, the inefficiency of repetitive data cleaning and processing tasks, and the cumbersomeness of visualisation tasks.
In particular, a lot of time is spent on the process of integrating and analysing data distributed across multiple sheets, which reduces the time available for deriving practical insights.
In addition, it is difficult to establish an automated process even though the same type of report is created every month, causing the person in charge to experience inefficiency as they have to repeat the same work from scratch every time.
The process of establishing a material ordering plan is the task that most clearly shows the limitations of Excel inventory management.
This task, which requires simultaneous consideration of complex variables such as the current inventory level of hundreds of items, safety stock standards, lead times, and minimum order quantities, is almost impossible to manage effectively using Excel.
In particular, the limitations of Excel become more apparent when you have to consider the different ordering cycles and delivery times for each material, as well as the production schedules of suppliers.
Incorrect ordering plans can lead to excessive or insufficient inventory, which directly leads to increased operating costs for the company. To minimise this risk, the people in charge are forced to spend excessive amounts of time.
The most dramatic change after the introduction of the AI inventory management system is the innovation of the data input process. Also, the time spent on data input can now be used for analysis and improvement activities to increase productivity.
Deepflow automatically collects and processes not only internal ERP data, but also a wide range of external environmental data, including 1,700 macroeconomic data, 6 million trend data, 100 industrial data, weather data, and industrial special event data.
This automation reduces the need for manual data entry, shortens work time, and minimises the possibility of human error.
In particular, 224 advanced AI prediction models, including the ensemble prediction model, can be used to accurately predict demand. Unlike traditional statistical models or general-purpose machine learning services, Deepflow's specialized model can analyse complex data patterns and accurately predict future demand to improve inventory management efficiency.
In addition, the user-friendly UI provides various information, such as forecast results, analysis of external environmental factors, and forecast of changes in inventory levels, in a visual manner. This allows users to easily understand and analyse data and use it for decision-making.
The AI system has made it possible to identify the inventory status in real time. All inbound and outbound movements and production inputs are reflected in the system in real time, which enables immediate response to urgent orders from the sales team or inventory inquiries from the production team.
In particular, even for products with complex BOM structures, AI automatically calculates the required materials and warns in advance of possible inventory shortages.
This has greatly reduced emergency situations such as production stoppages or urgent orders due to a shortage of materials, and has increased the predictability of inventory management. In addition, multiple departments can access the system at the same time, enabling real-time sharing of information and quick decision-making.
The introduction of the AI inventory management system has fundamentally solved the problem of data integrity. Problems such as formula errors and broken reference relationships have been completely eliminated through centralized database management, and all data changes are automatically recorded and traceable.
The system performs automatic verification of all incoming data and immediately generates a warning if any abnormal values or patterns are detected.
In addition, the rights management system has enabled the systematic management of data modification rights, preventing unintended data changes or losses. These changes have greatly improved the reliability of data-based decision-making and made it easier to track and resolve the causes of problems.
The AI system has solved the problem of discrepancies between physical inventory and inventory data on the system. Real-time inventory management linked to the barcode system, automatic inventory measurement through IoT sensors, and regular inventory verification support functions have enabled accurate synchronization of physical inventory and system data.
In particular, the function of AI learning past inventory fluctuation patterns to detect and alert abnormal inventory changes is greatly helping to prevent inventory discrepancies from occurring in advance.
This precise inventory management is increasing the reliability of production plans and contributing to reducing unnecessary safety stock.
The AI system has completely automated the process of analyzing inventory status and creating reports. Various analysis reports are automatically generated based on predefined templates, and key indicators can be monitored in real time through the dashboard.
AI analyses vast amounts of data to provide insights on inventory turnover, optimal inventory levels, and order timing. In particular, predictive analytics functions such as demand forecasting through time series analysis, forecasting the risk of inventory shortages, and overstock alerts have enabled proactive inventory management.
This has enabled managers to invest their time in deriving practical improvement measures rather than simply organizing data.
AI has dramatically improved the complex process of establishing ordering plans. The system comprehensively analyses various variables, such as past usage patterns for each item, lead times, supplier characteristics, and market conditions, to propose the optimal ordering plan.
The AI demand forecasting model provides highly accurate forecasts by taking into account seasonality, special circumstances, and market trends, and based on this, automated ordering proposals are made.
In particular, the supply chain risk analysis function allows us to detect potential supply and demand issues in advance and respond to them, which has greatly improved the reliability of inventory management.
With the help of this system, our staff can now spend more time on exceptional situation management and strategic decision-making.
The introduction of Deepflow has implications that go beyond simple changes to business processes.
The most notable change is that a culture of data-driven decision-making will be established throughout the company. Inventory management decisions, which used to rely on past experience and intuition, are now based on objective data and scientific analysis, which has enabled predictable and stable inventory management.
In particular, as various departments communicate based on the same data, the efficiency of interdepartmental collaboration has been greatly improved and enterprise-wide optimisation has become possible.
The work paradigm of inventory managers has also changed significantly. They are now able to focus on data analysis and strategic decision-making, rather than on simple repetitive data entry and checking.
In particular, it has become possible to establish inventory management strategies and take preemptive action against abnormal situations based on the insights provided by Deepflow. This has led to an improvement in the job satisfaction of the personnel in charge and has enabled them to develop a higher level of expertise.
The introduction of the Deepflow system will ultimately lead to an increase in the company's competitiveness. It has shown tangible results in various aspects, including optimisation of inventory costs, improved responsiveness to emergencies, and enhanced customer service levels.
In particular, in the current business environment where supply chain uncertainties are increasing, AI-based predictive inventory management has become an opportunity to take enterprise risk management to the next level.
The transition from Excel-based inventory management to AI systems is no longer an option, but a necessity. This is a key task that goes beyond simply streamlining work and is essential for accelerating digital transformation and securing future competitiveness.
Through the introduction of Deepflow, many companies are expected to continue to create greater value through continuous development and innovation.