A SYSTEM THAT INCREASES OPERATIONAL EFFICIENCY IN THE RETAIL SECTOR.

Abstract

The invention relates to a system that solves technical problems encountered in the retail sector using artificial intelligence and big data analytics, optimizing demand forecasting, inventory management, and dynamic pricing processes, thereby increasing operational efficiency and customer satisfaction.

Description

1 TARIFF A SYSTEM THAT INCREASES OPERATIONAL EFFICIENCY IN THE RETAIL SECTOR. Technical Area The invention combines artificial intelligence with analytics of variable retail sector parameters. a system that enables transformative dynamic pricing and user interaction based on 5 It is related to. The invention is particularly useful for solving technical problems encountered in the retail sector using artificial intelligence and... Solving demand forecasting, inventory management, and dynamics using big data analytics. By optimizing pricing processes, we can improve operational efficiency and customer satisfaction. It is related to a system that helps increase satisfaction. 10 State of the Art Today, data scarcity is a significant problem in demand forecasting algorithms. Historical sales data may not always be accurate and up-to-date. Incomplete or inaccurate data, This leads to inaccurate predictions. Furthermore, these models lack elasticity and are susceptible to sudden demand fluctuations. They cannot adapt quickly to changes. Seasonality and trends are not sufficiently understood. Algorithms that cannot model can make incorrect predictions. This can lead to excess inventory or... This leads to problems such as insufficiency, increasing costs and reducing customer satisfaction. negative effects. In inventory management systems, the lack of real-time data is a major problem. ERP These systems typically process end-of-day data, which enables real-time inventory management. This makes it difficult. Furthermore, data integration between different systems is often challenging, and data This can lead to inconsistencies. These problems can result from the mismanagement of stock levels and This leads to a decrease in operational efficiency. In pricing strategies, non-dynamic pricing is a major shortcoming. Fixed pricing is preferable. Pricing strategies, rapid response to changes in demand and competitive conditions 25 It cannot. Accurately measuring and managing the impact of promotional pricing. This is also difficult. This situation makes revenue optimization challenging and is due to dynamic market conditions. It prevents adaptation. Current retail practices and algorithms handle inventory management, demand forecasting, and It has several shortcomings in terms of pricing. Most existing solutions use data from 30 2 unable to analyze quickly and accurately enough and make real-time decisions. They struggle to capture this. Furthermore, these systems often lack the dynamics to handle customer interactions. They are somehow failing to manage it effectively. These situations include excess or shortage of inventory, technical problems such as incorrect pricing strategies and low customer satisfaction It creates. 5 In summary, current demand forecasting, inventory management, and pricing systems generally inadequate in real-time data analysis and adapting to sudden changes. These shortcomings remain, reducing operational efficiency in the retail sector and It leads to technical problems that negatively affect customer satisfaction. Artificial intelligence and Advanced technologies like big data analytics are critical to solving these problems. has. As a result of the research conducted on this subject, the numbered “Method and An application titled "apparatus for dynamic online pricing" was found. The system, a method for presenting a potential customer with dynamic pricing of a commodity and related to the device, processing real-time data using artificial intelligence methods 15 It does not perform this function; it updates at specific intervals. In conclusion, due to the negative aspects described above and the current solutions being the subject of discussion... Due to its shortcomings, an improvement is needed in the relevant technical field. It has been made. The purpose of this invention is 20 The invention was created by drawing inspiration from existing situations and overcoming the aforementioned drawbacks. It aims to solve the problem. The main purpose of the invention is to solve technical problems encountered in the retail sector using artificial intelligence. and solves demand forecasting, inventory management and dynamics using big data analytics. By optimizing pricing processes, operational efficiency and customer satisfaction. The goal is to create a system that increases customer satisfaction. Another purpose of the invention is to perform real-time data stream analysis using Spark Streaming. a system that does this and adapts quickly to instantaneous changes in demand This feature enables high accuracy in inventory management and demand forecasting. 3 Another aim of the invention is adaptive learning that works with machine learning algorithms. With their continuously updated and learning structures, their models are used in demand forecasting. The goal is to develop a system that provides high accuracy. These models are seasonal and trend-sensitive. It makes dynamic predictions, taking into account fundamental changes. Another aim of the invention is to implement demand-responsive pricing using multi-criteria decision analysis. The goal is to present a system that generates models. These algorithms determine product prices based on demand and inventory. It optimizes according to their levels, such as criteria like maximum turnover or maximum profitability. It provides dynamic pricing by taking this into account. Another aim of the invention is to enable users to interact with the platform naturally by using large language models. The goal is to create a system that enables interaction in a language. This feature allows users to interact in 10 languages. It allows for complex analyses to be performed through simple queries. Another objective of the invention is to process large amounts using Hadoop and cloud-based solutions. a system that enables the fast and efficient processing and storage of data These technologies enable scalable and flexible data management. To fulfill the purposes described above, the invention has 15 applications in the retail sector. It is a system that increases operational efficiency. Accordingly, the system;  By monitoring customer demand and stock levels in real time, we can provide analysis. detecting a sudden increase in sales of certain products at a particular store updating stock levels when necessary, adjusting the supply chain according to changes in demand. Data analysis that optimizes and records the analysis results in a database. 20 module,  When an increase or decrease in demand for a particular product is detected, the price is automatically adjusted. dynamic pricing module that allows the pricing to be increased or decreased accordingly,  Informing the user about layers, tables in the database and data analysis a retail assistant that informs the user about the results and a 25 chatbots allow users to interact with the system in natural language. NLP unit providing,  Provides all data analysis as a report for decision-makers and offers dynamic pricing. for the module, the user-defined ratio for the maximum profitability-revenue balance. According to the reporting unit, which provides dynamic recommendations, 30  providing the processing power necessary for the system to operate and the system's changing A cloud system that enables scaling of processing power according to usage demand. 4  converts data from different data sources into a standard structure and this data that enables the system to integrate with different external systems. integration unit It includes. The invention also provides a method for increasing operational efficiency in the retail sector. It also includes. Accordingly, the method is:  Raw data on customer demand, inventory levels and supply chain processes, store sales records, inventory status, supplier data, regional demand information, and Various data, such as external market data, are analyzed by a data analysis module to create a real 10 timely collection  The data collected by the aforementioned data analysis module is incorrect or incomplete. processes such as identifying and correcting information, and filtering out unnecessary information. by processing it and making it available for analysis,  The data analysis module detects a sudden increase in the sale of a specific product in a particular store. When it detects an increase, it automatically updates stock levels and supplies. optimizing its supply chain according to these changes,  The analysis results of the data analysis module are uploaded to a database by other modules. to save it in a way that can be used by,  A dynamic pricing module containing dynamic pricing algorithms 20 Dynamic pricing algorithms determine pricing based on demand and inventory levels. By staying between the lower and upper limits, it provides the user with price suggestions within those limits. presenting,  an NLP unit that informs the user about the layers in the database a 25 that informs the user about tables and data analysis results. Through a retail assistant and a chatbot, users can interact naturally with the platform. interacting in the language, and the aforementioned dynamic pricing module and data. learn about the analyses performed by the analysis modules ensuring,  A reporting unit should present all data analysis as a report for decision-makers. 30 and for the dynamic pricing module, a maximum profitability-revenue balance. It provides dynamic suggestions based on the rate specified by the user.  the processing power required for the system to operate and according to changing usage demand Scaling of processing power through a cloud system 35  data that converts data from different data sources into a standard structure the integration unit integrating the system with external systems It includes the steps involved in the process. The structural and characteristic features and all the advantages of the invention are given in the figures and 5 below. This becomes clearer thanks to the detailed explanation written with references to these figures. This will be understood as such, and therefore the evaluation will also take these forms and detailed explanations into account. This should be done taking that into consideration. Ways to Help Understand the Discovery Figure 1 shows a schematic representation of the system that is the subject of the invention. 10 Explanation of Part References 1. Data analysis module 2. Dynamic pricing module 3. NLP unit 4. Reporting unit 15 5. Cloud system 6. Data Integration Unit Detailed Description of Find In this detailed description, the preferred configurations of the system that is the subject of the invention are listed only. This is explained to facilitate a better understanding of the subject. 20 The invention is being developed with two fundamental analytical layers. One of these is the optimization layer. Demand forecasting, inventory optimization, and supply management are all enabled through the dynamic pricing module. It includes modules for inter-store transfers and unsold product analysis. The interaction layer is a retail assistant where user interaction takes place. It contains. 25 6 The invention leverages big data analytics and artificial intelligence to create dynamic solutions within the retail sector. Optimizes pricing and user interaction. Data analysis module (1), customer It monitors and analyzes demand and stock levels in real time. Dynamic pricing module (2), demand and stock with dynamic pricing algorithms While optimizing prices according to their levels, the NLP unit (3) enables users to interact with the platform 5 natural language interaction, dynamic pricing and data analysis modules It enables them to obtain information about the subject. All data analysis and dynamics for decision-makers. The pricing module sets the maximum profitability-revenue balance as defined by the user. Dynamic recommendations based on the rate are provided by the reporting unit. The system, the system's 10 that provides the necessary processing power for its operation and responds to the changing usage demands of the system. on a cloud system (5) that enables scaling of processing power according to It is positioned. The system processes data from different data sources in a standard way. It is integrated with external systems through the data integration unit (6) which transforms the structure. The detailed methods for the activities to be carried out are listed below; The data analysis module (1) belongs to the optimization layer mentioned above and the data 15 all processes from obtaining the data from the source to evaluating the results It includes. In the first stage, customer demands, stock levels and supply chain Raw data regarding processes includes store sales records, inventory status, and supplier data. It is collected from various sources such as regional demand information and external market data. Hadoop and Spark platforms enable the processing of large amounts of data quickly and efficiently. 20 This collected data is used to identify and correct erroneous or incomplete information, and to remove unnecessary information. It is processed through procedures such as filtering and then made available for analysis. For example, If discrepancies are found between stock levels, this data is corrected or removed from the analysis. This preprocessing process transforms the data into a coherent and analyzable structure. It enables conversion. 25 After the data preprocessing step is complete, real-time analysis is performed. Thanks to these analyses performed with tools like Spark Streaming, in a particular store When a sudden increase in sales of a particular product is detected, the system automatically adjusts the stock. It updates its levels and optimizes the supply chain according to these changes. Similarly Over time, by analyzing customer transactions and inventory movements, it is determined which products are suitable for which 30% of the market. It is determined that there is more demand in these regions. These analyses support regional stock management. optimizing strategies and ensuring a balanced distribution of products across different stores It contributes to the distribution of excess stock in a store. 7 This prevents stock shortages in other stores while this is happening. In general These processes, therefore, reduce inventory costs while increasing customer satisfaction. Inventory optimization using combinatorial optimization and stochastic modeling techniques. Demand levels, forecasting, and supplier management are optimized. Supplier relationship management. Processes are improved based on these analyses, and supply costs are reduced, resulting in procurement 5. The supply chain efficiency is increased. Supplier selection is done for the fastest and most cost-effective suppliers. Performance data is analyzed. Stock levels are continuously monitored through automated replenishment processes. This ensures that product stock levels are maintained at an optimal level. Additionally, product stock levels and demand forecasts are also monitored. Optimal decisions are made based on analyses of costs and supplier management. Recurrent neural network (RNN) and LSTM architectures in time series data analysis 10 While XGBoost, CatBoost, LightGBM are used effectively in structured data, Powerful demand forecasts are achieved through artificial learning models like NGBoost. In the data analysis module (1), special analyses for unsold products This has been carried out. Within this scope, RFM analysis, FSN analysis, and inventory turnover rate (stock) analysis were performed. turnover rate), association analysis, product lifecycle assessment, and 15 holiday days Many sub-studies have been conducted, such as those on its impact. The main purpose of these analyses is to examine the unsold... To identify the root causes of product-related problems and provide guidance to decision-makers accordingly. The analysis results include warnings that the products should not be restocked. When presenting these products, if it is determined that the lack of sales is a temporary situation, these products... It also provides feedback indicating that it can be restocked. This process, stock 20 to make effective decisions in management and to use resources more efficiently It contributes. Based on the root cause, it is recommended that products no longer be kept in stock. For the transfer to stores that still have sales, the stores must also be supplier stores. recognized by the system and routing work required in the transfer process This analysis is also useful in providing input to the inter-store transfer module. This is important. This way, unnecessary stock is avoided, and products that can no longer be sold are still in demand. They are directed to the stores. Finally, the analysis results are transferred to the database and used by other modules. It is integrated into the system in a way that can be used. These processes include dynamic pricing and It supports other modules, such as supply chain management, in making data-driven decisions and 30 It improves the overall performance of the system. 8 In the Data Analysis module (1), each sub-study feeds the dynamic pricing. It is geared towards this. However, inter-store transfer and supply management modules allow users to... It has been left to efficiency and cost savings. Supply management aimed at shortening lead times. In this module, decision-makers can adjust the system's factors such as suppliers' lead times and supply methods. They can run optimization by entering the desired details and 5 for product shipment. They can optimize the required lead times. Optimized results enhance the interaction layer. This allows the information to be shared with suppliers and the process to be carried out effectively. However The system can also utilize the delivery time suggested by the suppliers. Here The choice is up to the decision-makers. A similar option exists in the inter-store transfer module as well. The system is in place. The system operates according to specific scheduled dates (customer 10 stores (receivers) to which transfers will be made (weekly or monthly as required) and Identifying the source stores (providers) that will handle the transfer and one-off shipments using the most suitable routes. First, collect the transfers from the source stores upon departure, and then follow the most suitable routes. It is designed to distribute. Additional features will be developed here to meet customer requirements. The number of trips can be increased through these efforts, and the stores designated as source or recipient are 15. Additional rules can be added (X store should never run out of stock, priority buyer X). (It is a store.) Dynamic pricing module (2) derives from demand forecasts and inventory optimization. By integrating the collected data, market conditions and profitability rates for each product can be determined. It offers a comprehensive pricing strategy that takes into account all factors. Pricing algorithms, simple 20 Price recommendations based on demand and inventory levels, using regression models. It creates a decision-making process by presenting the user with alternative price options. A system has been designed that encourages participation in the process and creates a price scale. The system says, "If you apply this price, your sales speed and profitability will be like this, but otherwise..." with analysis-based recommendations such as "you can achieve these results if you choose a price" 25 It guides the user. Pricing is adjusted to include upper and lower limits, and these limits are based on the cost of the product and its lifecycle. The price limit is calculated taking into account the cycle and market conditions of the product. It is determined by adding 30% to the cost, and this guarantees minimum profitability. It prevents selling at a loss. The upper price limit is based on the product's life cycle analysis and 30 It varies depending on competitor prices. If the product's life cycle... If completed, the top price will be reduced by a certain percentage (e.g., 10%) from the competitor's price. It is calculated. If the product is in its mid-life stage, the system can add up to 15% more to the competitor's price. It can be removed, and if the product is only available in the system and not offered by other competitors, it can be removed. 9 The maximum price can be set by increasing the price by up to 20% above the competitor's price. Here... The goal is to maximize profitability on a product-by-product basis. Regression-based pricing algorithms determine demand by staying between these lower and upper limits. It takes its predictions into account and offers the user the most suitable price suggestions within these limits. The system also takes seasonal effects, such as holiday seasons, into account when pricing. They optimize their strategies. For example, increasing prices during periods of high demand both Optimize inventory turnover rate and profitability while setting prices during periods of low demand. By lowering the prices, the inventory is quickly depleted. With this approach, dynamic pricing goes beyond being just an automated process. It offers the user flexibility and control. The user can utilize the system's analysis-based 10 By following these recommendations, pricing decisions can be made more consciously. Thus, Market conditions, inventory optimization, and demand trends are evaluated together. The goal is to maximize revenue and improve inventory management efficiency. NLP unit (3), natural language processing and big language models with Chatbots or client It uses NLP technology for service automation. Firms like OpenAI have 15 The models it has developed (e.g., GPT series) have the ability to respond to queries. Chatbots have gained attention in the sector. Although chatbots are widely used in processes... Use cases that allow for an integrated approach to the entire system with Generative AI. It is not found in the retail sector. A manufacturer that integrates with major language models. Artificial intelligence techniques, automation of retail strategies and continuous improvement over 20 years It enables its development. NLP unit (3) uses advanced GenAI methods to provide the user with information about the layers. informing the user about the tables in the database and the results of the data analysis. It includes a retail assistant that acts as a proprietor, and a chatbot built with LLMs. The chatbot will be used by store managers as well as 25 retail suppliers. Aware of lead time optimization in areas such as the supply management module. Suppliers may also have certain authorizations in order to be able to do so. Authorization definition Savings belong to the stores. Access is via GTech's invention website using a username. This can be done after accessing the site with a password. The reporting unit, preferably characterized as a dashboard (4), provides decision-makers with 30 All data analysis is presented as a report, especially for dynamic pricing. according to the user-defined ratio for the maximum profitability-revenue balance for the platform It differs from other reporting systems in that it offers dynamic suggestions. This feature offers retailers an opportunity for profit management. It is being developed. The dashboard shows how each analysis works for our B2B customers. There are areas where one can see detailed results, even down to working with an algorithm. also, they can only examine profitability and dynamic pricing results. 5 Specialized pages are also offered. For product transfers, stores need to open this request through the system. The main locations where transferable stores, the most suitable receiving-transfer points and routes are found. The tables are saved to the database as outputs of the optimization layer. Users can view transfer information and track the transfer process through the interaction layer. They can initiate it themselves. The inter-store transfer module optimizes the entire transfer process. Even if they prepare it by prioritizing, security and approval processes are necessary. Therefore, it does not trigger the system itself. Cloud-based and microservice architecture offered with cloud system (5), the system It increases its scalability. Thanks to this, system performance can be improved by 15 in the face of increasing demand. It is protected. Cloud technologies increase efficiency by optimizing costs. By using cloud-based solutions and microservice architecture, the platform is flexible and It is designed to be scalable. During peak shopping periods (Black Friday) (during the campaign) the platform can process millions of transactions without its performance decreasing Amazon Web Services (AWS) is preferably used for this. 20 The invention can be easily integrated with existing ERP, CRM, and other business software. This Integration capability accelerates the adaptation process and aligns with existing business processes. In this context, different systems are provided on the data integration unit (6). APIs have been developed to enable data exchange between them. This increases the system's flexibility and It increases compatibility. ETL processes and API-based integration methods 25 By using this method, data from different data sources are brought together into a homogeneous structure. Sales data from an old ERP system is being transformed into a modern CRM system. It can be analyzed by integrating it with the system. In this way, customer behavior and Developing marketing strategies by identifying correlations among sales trends. is provided. 30

Claims

11 REQUESTS 1. It is a system that increases operational efficiency in the retail sector, and its features include:  By monitoring customer demand and stock levels in real time, we can provide analysis. detecting a sudden increase in sales of certain products at a particular store updating stock levels when required, adjusting the supply chain according to changes in demand. Data analysis that optimizes and records the analysis results in a database. module (1),  When an increase or decrease in demand for a particular product is detected, the price is automatically adjusted. dynamic pricing module that allows the price to be increased or decreased (2), 10  Informing the user about layers, tables in the database and data analysis a retail assistant that informs the user about the results and chatbots allow users to interact with the system in natural language. NLP unit providing (3),  Presents all data analysis as a report for decision-makers and provides dynamic pricing 15 the ratio determined by the user in the maximum profitability-turnover balance for module (2) reporting unit that provides dynamic suggestions (4),  providing the processing power necessary for the system to operate and the system's changing A cloud system that enables scaling of processing power according to usage demand. (5), 20  converts data from different data sources into a standard structure and this data that enables the system to integrate with different external systems. integration unit (6) It includes.

2. It is a method that increases operational efficiency in the retail sector, and its characteristic feature is;  Raw data on customer demand, inventory levels and supply chain processes, store sales records, inventory status, supplier data, regional demand information, and Various data such as external market data are realized by a data analysis module (1) collected in a timely manner, 30  The data collected by the mentioned data analysis module (1) is erroneous or incomplete processes such as identifying and correcting information, and filtering out unnecessary information. by processing it and making it available for analysis, 12  data analysis module (1) sudden increase in the sale of a specific product in a specific store When it detects an increase, it automatically updates stock levels and supplies. optimizing its supply chain according to these changes,  The analysis results of the data analysis module (1) are transferred to a database. saving it in a way that can be used by modules, 5  a dynamic pricing module that includes dynamic pricing algorithms (2) dynamic pricing algorithms based on demand and stock levels By staying between a lower and an upper limit, it provides the user with price suggestions within those limits. presenting,  an NLP unit (3), which informs the user about the layers, 10 in the database tables and data analysis results that inform the user about the information. Through a retail assistant and a chatbot, users can interact naturally with the platform. interacting in the language, and the aforementioned dynamic pricing module (2) and learn about the analyses performed by the data analysis modules (1) verification, 15  a reporting unit (4) reports all data analysis for decision-makers maximum profitability-turnover for offering and dynamic pricing module (2) In terms of balance, it offers dynamic suggestions based on the ratio determined by the user.  the processing power required for the system to operate and according to changing usage demand Scaling of processing power operations through a cloud system (5) 20 ensuring,  data that converts data from different data sources into a standard structure (6) the integration unit integrates the system with external systems It includes the steps of the process.

Images