‘Forecasting’ the future of demand scenario planning with Gen AI

In periods of high volatility, CPGs must quickly evaluate shifting market forces and consumer demand to make informed decisions about commercial and customer investments, capacity and inventory planning, and cash allocations. These decisions must be scaled across various product categories, industries, locations, and channels served by the business. Effective demand planning and forecasting are key to building adaptable and resilient supply chains for seamless fulfillment.

 

This blog delves into the limitations of traditional demand planning and forecasting methods and how generative AI can enhance these processes, making enterprises more agile and responsive than ever.

Demand planning and forecasting explained

Demand planning involves predicting future sales (i.e., future demand), helping retailers and suppliers gauge the volume of customers requiring their products. The goal is to ensure the right level of inventory and align execution capacity to deliver the right products to the right customers. Techniques like scenario planning and sensitivity analysis improve forecast accuracy by creating multiple hypothetical scenarios with different assumptions to predict future demand.

 

Similarly, demand forecasting helps companies make informed decisions about production, inventory management, pricing, and other critical operations. It provides vital information about future customer demand, which can help in minimizing waste, reducing costs, and increasing operational efficiency. Together, demand planning and forecasting lead to better financial performance by ensuring the right products are available in the right quantities at the right place.

Why demand planning is crucial for business

Operating decisions often hinge on current market demand for a given product and a company’s ability to satisfy that demand profitably. These decisions are backed by effective demand planning that requires collaboration among key stakeholders within and outside the organization to align supply with demand.

 

Demand planning is a critical business practice that positively affects the bottom line, while offering additional benefits such as:

 

 

The shortcomings of traditional demand planning

Traditional statistical demand forecasting generally relies on historical data spanning 24-26 months. Limited scope and inability to capture multiple factors lead to inaccurate predictions and missed opportunities. The limitations of traditional demand planning can bring down the efficiency and effectiveness of overall business operations.

 

Some of the drawbacks of traditional demand planning tools include:

 

• Lack of formal guidance on where trends and uncertainties originate
• Inability to identify the right scenarios to simulate and plan for
• Absence of strategic recommendations to mitigate risks in conflicting scenarios
• Increased costs due to poor information flow and forecasting practices across silos
• Reliance on historical data and linear forecasting models that fail to account for complex market dynamics or sudden demand changes
• Lack of agility to respond quickly to changing market conditions, hyper-competitive landscapes, and evolving consumer preferences

 

Relying on traditional demand forecasting methods can lead to missed opportunities, higher capital costs, excess inventory, waste, and increased risks of obsolescence or perished products—a significant concern in the fast-paced CPG industry.

Bridging the demand planning gap with generative AI

It is well-known that AI and ML can analyze structured and unstructured data from thousands of dynamic sources and real-time signals. AI-driven demand forecasting can significantly improve accuracy by modeling various scenarios, capturing assumptions, and anticipating trends, enabling companies to stay agile amidst market changes.

 

However, Gen AI takes demand planning to the next level. Rather than replacing existing AI and ML models, it enhances them.

 

• Gen AI creates a dynamic, real-time demand picture by analyzing multiple micro and macro-level indicators, offering an outside-in perspective.
• Acting as an additional layer of analytics, it augments existing approaches and pushes predictive capabilities to new heights. This allows planners to spend less time manually fine-tuning forecasts and focus on strategic decision-making and rapid responses to market shifts.
• Gen AI empowers planners with deeper insights that enable strategic decisions for driving growth and resilience.
• Gen AI can significantly enhance an organization’s capability to conduct robust contingency scenario planning faster and at a much lower cost than conventional processes.

Gen AI use-cases for demand planning and forecasting

Sigmoid’s Framework of a smart demand planning tool

At Sigmoid, we believe the most accurate forecasts come from a balanced approach that combines human expertise with advanced technology. While algorithms excel at identifying patterns in data, human insight is essential for interpreting elements that go beyond the data. Our strategic demand planning framework follows a ‘smart-touch’ philosophy, integrating both human and technological strengths to deliver optimal, precise forecasts in a dynamic environment.

Our machine learning engine generates an unbiased, accurate baseline forecast using market intelligence and a deep understanding of demand drivers. The generative AI interaction layer enables automated enrichment to ensure this baseline forecast is enriched with relevant information to achieve the desired forecast accuracy.

 

 

Generative AI can facilitate effective enrichment through the following capabilities:

 

• Synthesizing data business assumptions from various enterprise planning systems, business documents, unstructured datasets, and assumption repositories for future planning
• Unlocking market intelligence from business data by establishing relationships with measured data for a more holistic measurement of planning outcomes
• Incorporating diverse commercial inputs into the planning process, including sales data, promotion data, market trends and consumer behavior insights and allocating weightage to these drivers for refining strategies for specific functions, channels and markets
• Generating future scenarios and shortlisting them by assigning a high-medium-low likelihood of occurrence and determining outcomes for the prioritized demand scenarios
• Observing and indicating changes in data across time periods in the form of insights
• Optimizing forecasts based on historical changes and planner feedback
• Facilitating real-time calculations with granular insights for decision-making
• Establishing benchmark ranges for key metrics such as forecast accuracy and bias

 

Best practices for implementing GenAI in demand planning and forecasting

Before deploying and scaling Gen AI, it’s crucial to assess its feasibility, evaluate its true value, and determine its scalability within the organization to fully realize its benefits. Here are a few best practices for success in demand scenario planning and forecasting:

 

 

FAQs

1. What is the difference between demand planning, demand forecasting, and demand sensing?
   • Demand Forecasting involves predicting future demand based on historical data, market trends, and statistical models. It typically looks at long-term trends.
   • Demand Planning is a broader process that incorporates demand forecasting to align inventory, production, and resources to meet predicted demand while optimizing business operations.
   • Demand Sensing is a short-term approach that uses real-time data, such as current sales and market conditions, to adjust forecasts and improve responsiveness to sudden demand shifts.




2. Can demand planning and forecasting complement S&OP strategy?

   Demand planning is often conducted in the early stages of sales and operations planning (S&OP), a process for aligning supply with demand by coordinating sales planning and forecasting with production planning in a unified approach. Timely planning enables companies to make sure supply meets any expected demand fluctuations. It also informs the manufacturing process, while ensuring the capacity and availability of materials to meet projected production goals.




3. What are the types of demand forecasting?
   • Passive: The forecast incorporates historical and other data that the company collects, and projections are based on patterns and trends that have held up over time and may happen in the future.
   • Active: The active approach takes into account aggressive growth plans such as marketing or product development and also the general competitive environment of the industry, including the economic outlook, market growth projections, and more.
   • Short-term: Short-term demand forecasting looks at a small window of time to inform the day-to-day operations.
   • Long-term: Long-term demand forecasting is conducted for a period greater than a year, which helps to identify and plan for seasonality, annual patterns, and production capacity.
   • Internal (micro-level) forecasting: These forecasts use firm-level data including historical sales, past and current financial metrics and sales team projections, and customer data to predict demand for particular products and services.
   • External (macro-level) forecasting: Macro-level demand forecasting is useful to incorporate larger industry-led, economic, national trends and other pervasive factors into an organization’s planning and projections.




4. How are demand planning and supply planning different?

   Demand Planning focuses on predicting future customer demand to ensure the right amount of inventory, resources, and production are available. It aims to align business operations with customer needs.

   Supply Planning concentrates on managing the production, procurement, and delivery of goods or services. It ensures that sufficient supply is available to meet demand while optimizing costs, lead times, and capacity constraints.

    

5. What are some of the techniques or methods for demand forecasting?

   Many different methods can be used to forecast demand. While some businesses employ a simple single technique, other businesses may employ mixed-method approaches that combine large data collection efforts with machine learning techniques to determine forecasts. However, demand forecasting techniques may be broadly categorized into qualitative and quantitative approaches.

   Qualitative demand forecasting methods include Delphi method, salesforce composite forecast, expert opinion, market research/surveys, focus group customer interviews. Quantitative demand forecasting methods include moving averages, trend projection and simple regressions, seasonal averages, Econometric modeling (advanced regressions), gradient boosting and other machine learning tools.