The challenge
Sales are hard to forecast when what people buy and market conditions keep changing. Arnott's used traditional forecasting methods, and these limited how accurately it could predict demand. That affected how it managed stock and how it planned sales.
What was delivered
Our team was involved in a careful review of Arnott's internal business data, to check that it was ready to support AI. A forecasting model learns from past data, so its forecasts can only be as reliable as that data. This step came before anything else.
After the assessment and preparation, AI sales forecasting models were introduced to give Arnott's a clearer view of what the market would buy.
What it included
- A check of whether internal business data was ready for AI
- Data preparation before any modelling
- AI sales forecasting models
- Forecasts that inform inventory and sales decisions
Results
Arnott's data was assessed and prepared, ready for more advanced forecasting models. Forecasts became considerably more accurate, which helped with stock management and sales planning, and the business had forecasts it could act on when making decisions. Figures are under NDA.
Why it matters
This program dealt with the data before any model was built. If your organisation wants AI forecasts, first check whether your data is complete, consistent and organised well enough for a model to learn from. We look at that in our Apply AI work and in the free Reviews.