AI Stock Re-Estimation for Price-Drop Demand Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods fail to effectively re-estimate stock and simulate demand after a price drop in wholesale/retail products, lacking system-wide views and requiring extensive calculations, and do not account for updated forecasts post-price drops.
Innovation Solution
A system utilizing AI engines to process data packets, extract attributes, and re-estimate stock parameters, performing operations like price causal analysis, sales forecasting, and output data at distribution centers, enabling forecasting post-price drops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional methods are used for stock re-estimation after price drops, then the analysis can be traced with restrictive assumptions, but the methods fail to account for updated forecasts and require system-wide views with extensive calculations
Solution Approach 1:
An AI engine is introduced as an intermediary component between data input and forecast output. The AI engine receives data packets, extracts attributes, and performs re-estimation of stock parameters, thereby handling the complexity of system-wide analysis while providing accurate updated forecasts without requiring manual system-wide views
Solution Approach 2:
The patent replaces conventional mechanical calculation methods with an AI-based system. Instead of using restrictive assumptions and manual system-wide analysis, the AI engine automatically processes data packets, extracts relevant attributes, and generates re-estimated forecasts, eliminating the need for extensive manual calculations while improving forecast reliability
2Loss of information
If simulation approach is used for demand analysis, then comprehensive analysis can be performed, but many more calculations are required compared to analytic counterparts
Solution Approach 1:
The AI engine extracts only the essential attributes from data packets that are relevant for re-estimation, rather than processing all possible data. This selective extraction maintains information completeness for forecast accuracy while significantly reducing the calculation burden compared to full simulation approaches
Solution Approach 2:
The system changes from using fixed restrictive assumptions to dynamically extracted attributes. The AI engine identifies and extracts relevant parameters from incoming data packets, allowing the system to adapt to current conditions without requiring exhaustive simulation calculations, thereby improving both information completeness and calculation efficiency
3Ease of operation
If conventional stock policies are maintained, then existing processes can be followed, but re-estimation of updated forecasts after price drops cannot be performed
Solution Approach 1:
The system transitions from static conventional stock policies to a dynamic AI-driven approach. The AI engine continuously receives data packets, extracts current attributes, and re-estimates stock parameters in real-time, enabling the system to adapt to price drops and changing conditions while maintaining operational simplicity through automated processes
Solution Approach 2:
The AI engine performs self-service by automatically receiving data packets, extracting relevant attributes, and generating re-estimated forecasts without manual intervention. This maintains process simplicity while enabling the system to adapt to changing conditions, as the AI autonomously updates forecasts based on current data without requiring complex manual policy changes
Data Source
AI summary
Present disclosure generally relate to stock re-estimation, particularly relates to methods and systems for re-estimating stock and simulating demand, due to price drop in online/offline wholesale/retail products/appliances. System receives attribute data, business context data, price change data, historical sales data, store related data, inventory data, discount data, input plan data as input. System performs feature engineering on input data to extract data latent variables, calendar features, demographics data, derived variables, web extracted data. System performs operations such as price causal, sales forecast, Price Segment (PS) causal, and output data at DC level and determines delta change, multiplication factor, price segment distribution from output data at site level. System obtains input plan data and determined delta change, multiplication factor, price segment distribution from output data at site level to compute re-order plan and output what if analysis, multi-level forecasting, forecast for extended time, demand sensing, seasonality simulation, ABC classification, reorder plan.