Learn how one of the world’s oldest tea manufacturers partnered with Nihilent to deploy Resense, an AI/ML-based forecasting engine that improved accuracy across B2B, B2C, and D2C channels, reduced excess inventory, and cut stockouts for key tea categories.
Service
AI/ML Demand Forecasting, Trade Promotion Analytics
Vertical
Manufacturing B2C, CPG / FMCG
Region
Global (Multi-country, multi-channel footprint)
Tech Stack
Resense Forecasting Engine, Localized Feature Engineering, Performance Dashboards
The client is one of the oldest and best-known tea manufacturing and distribution companies, selling through B2B, B2C, and fast-growing D2C channels. Its portfolio spans everyday teas and premium infusions, each with distinct demand patterns and promotional behaviors.
01
Low forecast accuracy drove severe regional inventory imbalances and capital lock-in.
02
Frequent stockouts in key B2B and D2C channels hurt customer service and brand perception.
03
Excess stock in specific product lines tied up working capital and increased write-off risk.
04
Promotional impact was hard to isolate, making trade-spend ROI measurement unreliable.
05
Complex, lumpy, and intermittent demand patterns were poorly handled by legacy models.
Nihilent used Resense and a consulting-led approach to design a forecasting model tailored to tea demand and channel complexity.
Segmented AI Engine
Resense segmented demand from fast-moving everyday teas to intermittent premium infusions and applied fit-for-purpose AI/ML models to each segment.
Localized Feature Engineering
The platform ingested localized weather, promotional calendars, and inflation metrics to enrich baselines and better explain demand spikes.
Baseline vs. Promotion Separation
Resense systematically separated pure baseline demand from promotional uplift to clarify true demand and promo ROI.
Performance Dashboards
Live dashboards tracked bias, accuracy, seasonality shifts, and promo impact for demand planners and commercial teams.
High-Touch vs. Touchless Approach
A portfolio framework ensured planners focused only on SKUs where human judgment added most value, with Resense managing the rest in a touchless mode.
in forecast accuracy across all sales channels strengthened planning quality.
in excess inventory freed critical working capital previously trapped in overstocked regions.
reduction in high-velocity D2C and key B2B channels directly protected revenue and brand trust.
Majority of the portfolio in a touchless forecasting segment, improving planner productivity.
gained clear visibility into baseline vs promotional uplift, enabling sharper promo and pricing decisions.
For omnichannel FMCG brands, accurate forecasting means understanding both everyday demand and promotion-driven spikes across products, channels, and regions. By deploying Resense with segmented AI models, localized external drivers, and clear baseline/promo separation, the client turned noisy, fragmented demand signals into a unified forecasting engine that improves accuracy, unlocks working capital, and guides more confident commercial decisions.
