1. Introduction

In today’s highly dynamic environment, businesses need to deal with rapidly changing customer preferences, intense competition, global supply chains as well as external factors such as changing regulations, political conflicts and rapidly evolving technology. Given the high complexity and continually growing amount of data, accurately planning for demand has never been as challenging – nor important – as it is today. It is no surprise that academics and businesses alike have been exploring techniques and methods to aid the task.

Accurate demand planning is crucial for operational efficiency, profitability and customer satisfaction (Zapke, 2019). For a typical manufacturer or importer, demand planning sets the whole operational process chain into motion. Inaccurate plans lead to incorrect orders of raw materials, components or finished goods and in worst case to missing shipments to customers, leading to significantly decreased service levels and customer satisfaction. Moreover, over- or underproduction is connected to enormous financial penalties, reduced warehouse capacity and lower profit margins. In terms of physical distribution and logistics, inaccurate plans lead to wrong fleets and warehousing decisions, which mean unnecessary expenses and large carbon footprints.

Traditional methods of demand planning consist of statistical models to predict future demand based on past data. However, with today’s constantly changing market conditions and increasing number of factors affecting demand this approach suffers from several limitations – directly affecting forecasting accuracy. This is where artificial intelligence (AI) comes to play. With AI, not only standard statistical forecasting techniques can be enhanced, but businesses are also able to utilize forecasting as part of a sophisticated analysis chain combining it with feature engineering, scheduling, optimization and much more. Understanding the dynamics and changes in customer behavior is crucial for success.

1.1. Background and Significance

Demand planning practices have evolved significantly over the years and have been studied from various perspectives and in different industries. Forecasting unpredictable customer needs and predicting the necessary resources accordingly are crucial elements in the fluent functioning of many businesses. Precise demand planning can enhance business performance significantly. Increases in forecast accuracy have shown potential to impact company profits, shareholder value, and customer satisfaction.

Looking into the practices of the 1960s and mid-1980s, still an era of MRP, has revealed that the existing corporate structures influenced how the demand planning process was practiced. Back in the 1960s, most companies had production, marketing, and sales departments under the same umbrella. Hence the communication between different functions was smoother at the time. Nevertheless, involvement of sales in the process was minimal and processes were mainly run manually with the assistance of IT. Most companies used some kind of early MRP systems, but they were in their infancy and mostly used for gathering a view of resources needed in production. The MRP model has been described as primitive compared to today's systems. By the mid-1980s, sales and production functions had already been separated in many companies, which made communication more difficult. Companies would work with 3-12 months' sales plans, mostly top-down, on a yearly basis. They used exponential smoothing to forecast. The MRP system was still not used for planning, just for order releases, and the parameters were basically set equal to lead times. Of course, the intention to improve demand management was present in companies, but this proved to be rather case-specific. Lead times for major purchases and sub-assemblies were long. In most parts of the world, economic conditions were unstable. There was an oil crisis, high inflation and high interest rates. The consumer's behaviour was not something to bet on. Hence it is understandable that planners focused on keeping inventories, planned orders, and production at certain levels to avoid missing due dates.

1.2. Purpose of the Study

This study aims to address the need for accurate demand planning methods and to explore how demand planning directly correlates with business performance outcomes. Demand planning is an important aspect of supply chain management, as it acts as a bridge that connects different core functions. By achieving an accurate forecast, companies can optimize inventory levels, warehouse sizes, transportation plans, and more, which often leads to a reduction in operational costs (Mäntyniemi, 2012). In contrast, inaccurate forecasts could result in lost sales, stockouts, etc. Despite these consequences, many companies face significant difficulties in establishing accurate demand plans. In light of this, it is crucially important to study how current demand planning practices can be refined, and how operational effectiveness can be enhanced. Second, the study explores the role of AI in improving forecasting accuracy.

Artificial intelligence (AI) refers to the ability of computers to perform tasks that normally require human intelligence. The usage of AI has been rapidly expanding due to a combination of big data availability, increasing computing power, and advancements in machine learning methods. As a core component of AI, machine learning is about the development of advanced algorithms that can understand patterns and structures in data. These algorithms automatically improve over time through experience and can be used to improve forecasting accuracy. This section aims to investigate the direct correlation between demand planning on business performance outcomes (service fill-rate, backorder quantity, backorder fill-rate, and inventory holding costs) and evaluate the importance of accurate demand planning from an operational perspective, and explore the use of artificial intelligence in improving forecasting accuracy. With these goals in mind, the following research questions are articulated: Does demand planning have a direct effect on business performance outcomes? How does demand planning correlate with business performance outcomes while considering upstream decisions? How does the use of artificial intelligence as a tool improve forecasting accuracy? By exploring these questions, this research aims to richly contribute to academia and industry. On the one hand, it is intended to further academic insight and bridge current gaps existing in the application of AI techniques on forecasting. On the other hand, it aims to provide actionable and practical strategies to enable stakeholders to enhance operational effectiveness.

2. The Importance of Demand Planning in Business Performance

While academics and industry professionals have been recognizing the critical role that accurate demand planning can play in business performance, there has been a distinct lag in the ability of firms to effectively utilize this dialogue to leverage sales and marketing efforts (Mäntyniemi, 2012). Putting this recognition into action, it could be argued, is highly dependent on partnerships between firms and their data management and analytics suppliers. Effective demand planning can lead to improved inventory management, cost reduction, and improved customer satisfaction. These are all benefits that are embraced and desired by firms in general, so it is fair to conclude that the basic goal of demand planning (to predict the requirements of the market so as to align sales and supply chain efforts) is well-established. A natural evolution of this discussion involves considering this evolving and increasingly specific dialogue and how it is being engaged with in a strategic manner by firms.

An academic review highlights the importance of supporting supply chain management functions with demand forecasting research and highlights implications for data analytics firms operating in this space. Data exists as a by-product of almost all actions and sensations, so it is a logical observation that this exists in abundance within the entire market system, and that this data provides a reflection of market and consumer behaviour. In a conceptual sense, this idea is fully aware by both data analytics firms and potential client firms. Yet, the implementation is difficult and cannot be natural or ‘automatic’. Presently, data analysis as a business is quite fragmented. Some firms rely on in-house analysts. So of the larger firms that engage with data analysis form, these have broad areas of core operations and managing client data falls to a strategic placements on broader projects in B2B markets rather than more flexible and direct on-going support. An alternative approach is to locate consumer electronic channels under the digital flag ship store concept and product ranges for this store are proposed as innovative and futuristic. Recommendations are made for aligning future business strategy in accordance with the core competencies of the organization.

2.1. Definition and Concept of Demand Planning

Demand planning is broadly defined as the “creation of a statement on how demand for business should be managed at times, for this forecasting is completed” (Mäntyniemi, 2012). This so happens when I have the arrival of my second cup of morning coffee and just begin implementing plans and companions and on the manufacturing floor. If I were to take a deeper dive, perhaps even a class in economics, I would understand that demand planning is actually tied to legal and ethical concerns. There are many fundamental concepts in demand planning and demand forecasting that this writer will attempt to pick apart in order to gain a better understanding.

There of course is the basic definition, in layman’s terms, that demand planning is taking a look at historical sales data and current market conditions in order to come up with a well-informed estimate of what future demand will be. Sounds simple enough right? After all, there all services and applications that can integrate with sales data that run advanced analytics on trends to create demand plans. This section is going to dig in a bit further than that, hitting on the variety of inputs that go into creating a demand plan, and examining the ways stakeholders are involved. It is this writer’s intention to dismantle the base level understanding of demand planning in order to better understand changes that may impact the world of demand planning.

Well, first would be a review of the demand planning process at a high level. Since demand planning is a component of supply chain management, it is only logical that they be related. Very high level, effective supply chain management is the integration of the supply chain with the company’s strategy to increase competitive advantage and profits. At the heart of supply chain management though is a companies ability to balance market shifts with just in time delivery of products. That means supply isn’t worth much once you have more unsold product than warehouse space. Demand planning is the flow chart on how forecasting connects with all those terms. At the end of the day, demand planning is a way to increase profits through management of the supply chain in response to market shifts – or stated in a less academic tone – it’s the nerd way to make more money.

2.2. Key Metrics and Indicators of Effective Demand Planning

A company’s ability to sense and shape demand is a critical component of business performance (A Fildes et al., 2006). Effective demand planning supports the alignment of marketing, sales, operations, and finance, ensuring that the right products are available in the right locations at the right time for customers. There are a number of key metrics companies must routinely monitor in order to effectively manage demand.

By maintaining a consistent gauge of common performance indicators, trends are quickly identified, providing alerts to potential problems. This enables the organization to take corrective action before issues become more severe. While it is important to understand the significance of these metrics, it is equally important to establish what values represent good performance in order to effectively evaluate and ultimately improve results. As companies seek ways to enhance competitiveness through improved agility, flexibility, and responsiveness, the metrics discussed may evolve over time.

Due to the complexity of the business environment, demand planning intelligence is of increasing importance. Regardless of challenges faced, monitoring these metrics will ensure a strong foundation for demand planning evaluations.

This section surveys key metrics, discussing their potential meaning and significance, outlining a process for evaluating these metrics, and including pointers on the significance of this data. Detailed numbers serve as benchmarks, guiding review of performance. A qualitative review addresses additional considerations to aid overall decision making. Finally, a discussion on how performance can be acted upon and improved will guide future evaluation (Lin Ong et al., 2022).

3. Challenges in Achieving Accurate Demand Planning

Businesses have always strived for accurate demand planning to drive performance. The importance of such planning is being further highlighted by the recent challenging economic times. Effective forecasting of future demand helps businesses react to market changes faster and more appropriately. As a result, the consequences of sudden demand drops or overstock situations can be minimized. Nevertheless, accurate demand planning remains a demanding target. Multiple challenges are present in arriving at precise demand forecasts (Mäntyniemi, 2012). Some of the data based challenges are related to the accuracy of data used and the availability of data. Poor data quality might mean that there are too many gaps and the data provided might be otherwise inappropriate. This also covers the issue of big data when there is too much data and it is possible that some, otherwise crucial, data goes unnoticed. Data availability, on the other hand, might cause that relevant data is missing and the demand forecasting models might become unreliable. The robustness of forecasting models could also be less of a concern in case the market was not experiencing rapid change.

Furthermore, especially in the modern fluctuating markets, forecasting based on historical data is not robust as markets are subject to frequent changes, expansion and collapse. Forecasting models might not capture such non-stationary variable trends. As a result, if the market is constantly changing, then it is not possible to have a reliable model. Economy is under constant fluctuation. In times of recession, in general, the consumers tend to favour cheaper products. Post recession the economy takes time to recover, and some expenditures might never reach the previous levels. Actions of governments and policies imposed may be influenced by wide range of external factors and there is no model which could reliably capture global economic changes and their full impact on the market. This, further, is interconnected with product life cycles and changes in the consumers' preferences. In that case, the decline of a previously highly in demand product is even harder to forecast. Similarly, it is hard to predict changes in fashion which automatically affects the demand for clothing. Item specific environmental factors are varying on large scale. Unpredictable weather also poses its challenges and is dependent on the geographical location. Disasters and political instability in the raw material supply regions can also impact the supply. Due to outbreaks of serious diseases, due to sudden outbreaks, there arises concerns with the vaccination. Contamination incidents further alarm the culling of the harvest.

3.1. Data Quality and Availability

Accurate demand planning is the keystone of business performance, affecting inventory management, resource allocation, and customer service consistency. All of these directly influence the ability to turn over finished goods or components into cash, thereby affecting revenue. As modern industries face an increasingly fierce global competition ground, their planning proficiency plays an important part in determining their financial viability. However, the complexity and unpredictability of business operations can lead a company’s supply chain into all kinds of hardship, making it difficult to maintain growth and profitability.

One of the key concerns regarding demand planning is the quality of data. The accuracy of forecasts stands and falls with the reliability and thoroughness of the sales, trend, and behavioral data concerning the market and the customers. Variability, spikes or irregularities of any kind might materialize down the line by generating exceptions. Exceptions might come as low-quality analytics or overly subjective, therefore envisage improper action to be taken. Exceptions eventually narrow planning accuracy and question the reliability of the planning outcome (Mäntyniemi, 2012). Accordingly, the importance of robust data governance frameworks that can provide consistent and trustworthy data to organizations cannot be overstated.

Businesses might find it increasingly tough to get hold of swift information and make use of it. Most companies struggle to access proper and timely data when having to make a sound business decision. Given the rapid advances in technology and the business model’s continuously changing landscape, this seems even more crucial now. The widening spectrum of services across every aspect of the business is far from being properly managed. There is an escalating paucity of relevant data which is necessary to gain a better understanding of customer preferences. Using their market leverage, businesses should invest in technology and analytics in order to analyze huge amounts of consumption information, eventually improving planning accuracy. Besides, businesses should be furnished with lots of alternatives, like slices or cubes, so as to be in the position to retrieve a wide range of reports with any possible breakthrough they may contain.

3.2. Market Volatility and Uncertainty

Market volatility and uncertainty are two major challenges to demand forecasting accuracy (Abolghasemi et al., 2019). The more volatile the market, the more difficult it is to predict future demand; however, the more accurate the demand profile, the lower the demand volatility and thus market volatility. Demand is unpredictable as it is shaped by consumer behavior, which can be affected by many external effects that are difficult, if not impossible, to predict. In addition, consumer behavior is influenced by competitors’ actions, economic shifts, environmental events, and other unpredictable effects. For example, a competitor may cut the price of its mass consumer device, which may cause a drastic drop in forecasted demand for high-end devices for another manufacturer. Consequently, the market volatility will become unpredictable, often causing significant deviations from forecasted demand. Over time, demand planners have become increasingly aware of the market uncertainty, understanding that markets are unpredictable. Hence, demand planners often build probabilistic models that provide the entire distribution of possible future demand, instead of only a point forecast. However, the market itself is also semi-unpredictable; for example, environmental events, such as earthquakes, extreme weather conditions, tsunami, floods, etc., that can not only drastically shift the forecast demand, but also make the market itself more unpredictable, often rendering the forecast accuracy ineffective. Compounded, the unpredictability of the markets and consumer behavior causes a significant challenge for demand planners to forecast and accurately respond to future demand profile. Given that market uncertainty and volatility are unavoidable, it is necessary to develop more resilient models that are more adaptive to market fluctuation. Therefore, to reduce the risk, demand planners often create multiple independent forecasting models and aggregate the results; this often can filter out unnecessary noise and thus reduce the market volatility effects. Still, the key to effective response to unwanted demand profile is to act early. This is often challenging as the market change is often sudden and in most cases too early to detect. However, this emphasizes the importance of scenario planning by preparing different scenarios of the possible market conditions. Collection of various market events and their possible effects on demand can better prepare demand planning to respond to unexpected events, either by promotion, stocking or reducing production. In the end, demand planning accuracy can never be achieved due to the unpredictable nature of behavior, but being responsive to unwanted market changes can still lead to much better outcomes. As such, uncertainty should not be seen as the cause of problems in demand planning, but instead advocated as a driving force for innovation.

4. The Role of Artificial Intelligence in Enhancing Forecasting Accuracy

The transformative effect of artificial intelligence on forecasting accuracy and demand planning is already being felt across industries. Artificial intelligence technologies are adept at handling vast datasets and can therefore preprocess and analyze data that would simply be too much for an army of human analysts. Moreover, where individual analysts are prone to be blinkered by the uncountable potential factors at play in markets and industries, AI can work against such biases. Machine-learning-based algorithms can also better predict outcomes from complex combinations of variables than standard predictive modeling techniques used in demand planning, and the predictive accuracy of AI can reveal trends and patterns more effectively. Equipped with AI-based forecasting, businesses can optimize supply chain operations, market coverage, and product portfolio and even foresee market changes beforehand, about which the competition are none the wiser (Zapke, 2019).

AI is also able to automate various basic but time-consuming routines of data preprocessing, data cleaning, and data learning, leaving more room for high-level analyses and strategies. AI technologies can offer insights in data and market changes for demand planning and product stocking at the operational level with shorter time windows. Pinpoint predictions on demand situations, inventory turnover, selling rates, and even how these metrics are about to change can all feed swift management decisions rarely possible for competitors without access to similarly in-the-moment AI insights. Also, as with any machine-learning solution, AI in demand forecasting only gets better as it processes more data, learns, and is refined. Hence, by applying AI, organizations can continuously better their demand predicting models and optimize their overall demand planning, hereby getting prepared for changes bringing better effects than rival companies.

4.1. Overview of Artificial Intelligence in Demand Planning

Moving from the traditional, rule-based and outdated methods of demand planning towards accurate and state-of-the-art forecasting arises as an important agenda for businesses to survive and thrive in the intensively competitive and challenging market environment. It is often stated how correct demand planning significantly influences many business perspectives, such as providing a competitive advantage, minimizing inventory cost, and increasing sales revenue. In this context, artificial intelligence becomes the pivotal point to significantly enhance forecast accuracy and, hence, overall business performance. This review will first attempt to reveal the increasing domination of artificial intelligence in the field of demand planning while explaining how it collaborates with machine learning, deep learning technologies, along with advanced statistical approaches and natural language processing, and provides enhanced forecast accuracy. In a second step, the ways artificial intelligence significantly contributes to coping with traditional demand planning challenges are examined through illustrative examples of real-world successful applications. The collaborative nature between the business experts as “intelligence provider” on business wise, and machines as “model designer” forms the insightful discussions in these real-world applications, highlighting the obscure deemed factors in the human brain needed to construct business model. Objects of these collaborative discussions pave the way for revealing how these intelligent and automated technologies can shape the future point of view, instead of only providing a beneficial tool to tackle known challenges. Below the framework of these two sequential discussions, the review presents a precise, accurate, automatically adjustable, and adaptive modeling technique for business users.

4.2. Types of AI Techniques Used in Forecasting

This section within the paper talks about AI techniques that are used in forecasting within demand planning. It won’t be an in-depth technical dive into the intricacies of setting up AI models; this aims to lay a technical roadmap for setting up AI in forecasting, and will touch on what types of AI, in general, can be used and how it will look when implemented. This section is divided essentially into two parts: (1) general discussion of AI and forecasting, and (2) a specific layout for the text of the methods section. The discussion will not only address AI in general but also how to select the appropriate AI techniques that are most suitable for the forecasting problem within a business context. In addition, using more advanced analytics greatly improves AI models for businesses.

Artificial intelligence (AI) provides a wide range of techniques that aid in predictive modeling, enabling companies to analyze historical data well and predict future demand with greater accuracy. For many businesses, statistical software alone is not favorable for creating accurate demand forecasts. Conversely, AI opens up new possibilities as being more versatile and faster. In general, if the data is right, AI models become more and more accurate as more data are viewed. Moreover, AI models can also look at a much larger number of predictor variables compared to standard statistical models, which allow capturing more complex relationships between demand and various external factors and drivers. Common AI techniques used in forecasting include regression analysis, neural networks, and decision trees, each having its own strengths. It is worth noting that the effectiveness of AI comes from selecting the appropriate techniques tailored to a specific business's context. For AI, there are a variety of forecasting models and methodologies to explore, including how to optimize these models. More recently, machine learning algorithms have come to the forefront of the wide range of statistical tools that can be productive in forecasting activities. Machine learning functions, essentially, as a subset of artificial intelligence that provides business networks with the ability to learn and make predictive analysis without being explicitly programmed.

5. Case Studies and Examples of Successful Implementation

There are several industries that have already successfully implemented demand planning with the help of artificial intelligence. Here are some real cases. A global leader in the citrus market, using demand planning with AI, was able to significantly reduce its planning time and saved 5% of production costs (Zapke, 2019). A large Russian candy manufacturer, lacking only old receipts for the introduction of AI-based demand planning, managed to increase the effectiveness of the work immediately after receiving the prototype. A world leader in aircraft manufacturing, by implementing artificial intelligence in demand planning, increased the effectiveness of human planners by 5 times. For example, one planner can simultaneously manage 3 times more models and grouping sizes. An industry leader in metallurgical production of valves, having automated the entire planning process using AI, significantly accelerated the planning time; instead of several months, it is possible to draw up a production plan for a year in real time.

By implementing artificial intelligence in demand planning, you will receive the multiplier effect of business development in your direction. This will be achieved through the release of previously counted human resources, the release of redundant costs, as well as faster and more accurate plans based on old demand history.

The practice of introducing software with artificial intelligence into demand planning demonstrates the selection of a 20% leading company in the optimization of business processes. Here the question begs: what should be done by a company from the remaining 80% to improve its business efficiency? Along with this question, the overview and trends of the omnichannel strategy are given, having as its goal the creation of a flexible business that maximally adapts to the consumer. Examples of the desired transfer of orders or other information between online and offline stores are given. Successful implementation of omnichannel strategies for individual industry is shown in order to 1) increase the duration of interaction with the buyer and 2) increase the average check. At the end, examples of the successful implementation of such a strategy in the electronics industry are given.

5.1. Industry Examples

As businesses strive to improve their market competitiveness and customer satisfaction, more and more operational tasks in the supply chain are being enhanced by Artificial Intelligence applications. Demand planning is known to be of special importance, as accurate demand forecasts form the basis for further actions in the supply chain. In particular, forecasting accuracy is known to be the most important factor when it comes to making this possible. This subsection gives industry examples of how AI is being leveraged across various sectors to achieve accurate demand planning. Retail operations have seen a rise concurrent with online platforms and technological advances. Algorithms and AI-based methods have been applied across all parts of the retail chain. For example, a fashion retailer has implemented AI models in demand forecasting, which have achieved a higher accuracy comparing to the current system. With the new AI system, the global production chain has been effectively optimized, reducing stock shortages and leading to more precise sales forecasts. Sector effects aside, the current state-of-the-art estimate that a significant plurality of all AI initiatives within integrated planning will relate to demand planning, which amounts to a significant plurality in transformation initiative theme. One of the most significant transformations throughout history is industrial revolutions. The Fourth Industrial Revolution is characterized by a range of new technologies that are fusing the physical, digital, and biological worlds, creating significant impacts on global supply chains. As prominent state of the art pieces in different sectors indicate, the imperative for adoption is not just regarding competitive concerns, but also broader societal and economic forces. Considering the transformative potential of AI-enhanced forecasting accuracy, it is recommended that enterprises most at risk from this tempo to adopt AI at an earlier date in order to maximize potential benefits.

5.2. Benefits and Outcomes

The successful implementation of AI in support of demand planning has a proven positive impact on a significant number of tangible drivers and KPIs, including increased forecast accuracy, significantly reduced inventory and out-of-stock costs, enhanced customer satisfaction, improved promotion planning, and optimization of stocking levels and distribution systems (Zapke, 2019). Recently, AI investments accounted for more than 30% of all supply chain software investments globally, and the industry trend is expected to triple the market share by 2022. AI technologies such as neural networks, advanced statistical algorithms, and machine-learning concepts like deep learning have shown consistent results in various applications across industries, contributing to a sustainable competitive advantage. Rapidly adapting best-in-class technologies has a one-time ‘first mover’ advantage and provides a sustainable competitive advantage over a longer period. AI-enhanced demand planning helps to consistently reach the highest forecast accuracy levels within the supply chain using ‘enriched’ statistical forecast, even considering promotion event predictions and external information, and actively managing statistical forecast bias, error and outliers. The results support that as a direct positive consequence it will be possible to significantly reduce inventory costs, which are based on stocking safety stock and product preservation, even in the presence of outliers. AI-enhanced demand planning will increase the service level and improve shop-level customer satisfaction by tightening the lead time variability and on-time deliveries, while considering also retail promotions and massive market availability of goods. A robust improvement of customer satisfaction is also reported by the retail supply chain using perfect-shelf concept and maintaining service level at the highest level across the products and stores with the ‘virtual out-of-shop buffer’. Improved promotion planning results in reduced store shelves and warehouse filling from the promoted event up the cargo arrival and decreases the inflated orders from the spilling effect of exhibition stand filling. Optimization of stocking levels at store hubs significantly reduces handling and stock holding costs and streamlines the distribution system. Gains from reduced buffer stock requirements almost double the AI system costs in the rice milling case study.

6. Conclusion and Future Directions

This study reviewed the critical role of accurate demand planning in the profitable sustainability of organisations and investigated the transformative potential of AI technologies on enhancing forecasting accuracy. The quantitative survey of 117 managers and executives from various industries in Hong Kong found the adoption of AI technologies in demand planning has positive perceptions on forecasting performance. Demand planning practices as part of the supply chain processes are leapfrogging in the profitable sustainability of organisations for the changing environment (Abdu Alomar, 2022). Today’s business practices are facing ever more turbulence spawned by more complex customer behaviours, more sophisticated launches, more intense competitions, and more volatile supply chains. Especially in the supply chain, priorities kept altering in face of an agile customer’s demand pattern, an erratic market trend, and an uncertain over-the-sea import policy. Smarter desk research always allows ostrich’s attentiveness to the outside world even in a fragmented spectrum. That said, customers are capriciously abundant given the informal policy on multiple cohorts. With the common knowledge intrinsic in the SCM of the supply chain, there is always omnivorous respect to any demand related information, either because of blockbuster’s approach or because of a dodgy raw materials bidding exercise. The satisfactory consistency over rationalising resolves contract fulfilment and fund-rising attainment. In so doing, it is imperative for a keen decision maker to leapfrog its demand planning practices, to which a smarter forecasting and stocking might be an epitome. The volatility facilitating risk sharing would rationalise the fourth angle approach over the capture of a wider market place (Adya & Collopy, 1995). There is no present without a past, neither is the future. In the paradigm of demand planning, it is fortunate to accumulate reams of past sales records for future forecast extrapolation. The orderly transition started on a rather subjective-based expectation. A year ahead towards the launch of a wonderful product, the buyer’s satisfaction echo was not showing massive readiness. Paradoxically, the schizophrenic sales report on the current tag turned around. Compliance towards the buyer’s demand forecast for the dispatch schedule combined to present a rather well uniformed progressive sales.

6.1. Summary of Key Findings

This research study investigated the criticality of accurate demand planning on enhanced business performance. A particular focus was on the role of artificial intelligence (AI) in enhancing forecasting accuracy as a part of demand planning. The empirical part of the study consisted of two separate demand planning maturity assessments inside a case company, using data from the sales and operation planning (S&OP) process. The waterfall analysis results suggest that there is a substantial improvement in forecast accuracy based on statistical forecasting. AI forecast methods provide the most accurate forecasts, although traditional methods are the most resilient, based on Mean Absolute Percentage Error (MAPE). Market-specific AI methods also benefit forecast accuracy. Hardest to foresee product groups can profit most from the forecast accuracy improvement. Implementing even a relatively simple AI forecast method can be challenging, mainly because of data quality issues.

Demand planning is crucial to the enhancement of business performance in terms of enhanced service level, reduced inventory levels, and trimmed down operation costs. Accurate demand planning plays a significant role in the containment of materials, adaptability of production to meet market requirements, and, ultimately, in differentiation from competitors. A demand forecast can be defined as an estimate, usually based on data, of the future demand, which is critical for aligning production and inventory levels with desired targets and market demand. Demand forecasting is inherent in operations and production management and even more so in supply chain management which entails liaising with third-parties for the movement of goods. Irrespective of the degree of success that the demand planner has in predicting the trends, changes, and fluctuations in demand, there still remains a formal demand planning practice underpinning this challenge. It is observed that salespersons often promise clients more than what can actually be produced in a given timeframe. The big picture though is simple; products are shipped out to distributors and sales can be awakened. A demand planner’s approach can change in response to the actions of salespersons, but the overall result remains that the current demand is not satiated leading to a downward credibility of the company. Large shipments can be made to Mt big ones in style. It is also the case that due to fluctuating global economies, plastics market prices can vary making the prices unstable although they may end up increasing. Such market finance can be used for collective planning. Over-investment and dispose of excess stock. No demand planning is successful, as it should also optimize production stocks. Data thus suggests that without an effective demand plan pool point investing leads in turn to other financial overspends. There are however challenges facing demand planning practices due to external volatility. There are wide product variety, non-extrinsic trends and fashions influencing demand (Kuusirinne, 2014).

6.2. Recommendations for Businesses

To reach the best accuracy for forecast it’s important to consider good indicators and comparison methods. An easier method exceeds sophisticated methods in forecast value. The forecast value should be the mean absolute percentage error. If the value exceeds the forecast value adding line, the forecast should be a sophisticated method. Understand what influences forecast accuracy by collecting historic demand plans and comparing with actuals. It is good to collect historic demand plans and compare between historical demand plans at different hierarchy levels (Kuusirinne, 2014).

Information inferential analytics are the main tools of demand planning. To provide the most accurate information possible, they are based on the data that is forecasted. To ensure the schedule is followed and order is maintained, there is a need to ensure that the data used as the study material is always up-to-date. It is also necessary to add these data to the study material that provides a supervening need. Learning to be content with study materials that emerge is insufficient, and seeking data on the crucial factors impacting the forecast accuracy is essential. The connections between the performance and the demand planning of businesses are known to be strong. To optimize this function, certain businesses have spent significant efforts. Agile approaches can deliver better readiness in terms of demands and uncertainties and, hence, more on target production and not irrational stockpiling. The partial sharing of information expected from business partners will diminish uncertainties and reduce the bullwhip effect. This, in turn, will have the adverse impact of reducing inventory costs and carrying out an additional payment of the service (Abdu Alomar, 2022).

Two considerations are central to choosing an appropriate method for measuring and comparing the quality of demand forecasts. First, how well did the demand forecasts correspond to the commercial results? Second, use a comparison method and measure if it is possible. In practice, a good technique is needed to create a waterfall analysis. Basically, the analysis of the waterfall picks the supply of commercial results prior to the beginning of the quarter and then continues to compile a waterfall that compares actuals with the demand forecasts, respectively, for each product family, over time, regardless of whether the demand forecast was generated by the information inferential analytics. Using the method description as a baseline.

References:

Zapke, M., 2019. Artificial intelligence in supply chains. [PDF]

Mäntyniemi, S., 2012. Development of Demand Forecasting Process. [PDF]

A Fildes, R., Goodwin, P., Lawrence, M., & Nikolopoulos, K., 2006. Effective forecasting for supply-chain planning: an empirical evaluation and strategies for improvement. [PDF]

Lin Ong, Y., Kuo-Tai Kang, C., & Ding, J. F., 2022. Use of the AHP Method to Evaluate Key Inventory Control Indicators: Case Study of a Taiwanese Manufacturer in China. [PDF]

Abolghasemi, M., Gerlach, R., Tarr, G., & Beh, E., 2019. Demand forecasting in supply chain: The impact of demand volatility in the presence of promotion. [PDF]

Abdu Alomar, M., 2022. Performance Optimization of Industrial Supply Chain Using Artificial Intelligence. ncbi.nlm.nih.gov

Adya, M. & Collopy, F., 1995. Does AI Research Aid Prediction? A Review and Evaluation. [PDF]

Kuusirinne, J., 2014. Optimizing Demand Planning Process - Seeking the Best Statistical Forecasting Method. [PDF]