Six AI-powered supply chain products. Explore each — see the problems they solve, the dashboards they power, and the results they deliver.
Ensemble ML demand forecasting using XGBoost, LightGBM, LSTM, and Prophet — with automated bias correction, demand sensing, tracking signal alerts, and uncertainty quantification. Replaces Excel-based forecasting with a system that learns and improves every cycle.
Request a Demo → Talk to UsA national FMCG distributor was running monthly forecasts on Excel for 14,000 SKUs across 8 distribution centres. MAPE averaged 38%, with undetected over-forecast bias of 15% on seasonal products causing $2.4M in annual excess inventory.
Mathnal deployed the SC Forecasting Engine with XGBoost + Prophet ensemble, external signal integration (weather, festivals, POS), and automated tracking signal monitoring. Within 4 months, MAPE dropped to 19%, bias was corrected to within ±3%, and safety stock was reduced by 22% — freeing $2.4M in working capital.
Automatically tests 6+ models per SKU and selects the best performer based on out-of-sample RMSE.
Integrates weather, POS, Google Trends, and economic indicators for short-horizon correction.
Tracking signal, CFE, and MPE monitored weekly. Auto-alerts when |TS| > 4.
Measures whether each process step improves or degrades accuracy. Eliminates waste.
Prediction intervals at 80% and 95% confidence for risk-aware safety stock calculation.
Built on scikit-learn, statsmodels, PyTorch. Deployable on AWS, Azure, or on-premise.
Dynamic safety stock using simulation, Bayesian methods, and multi-echelon modelling. Calculates optimal buffers at any service level — and tells you exactly how much working capital you can free.
Request a Demo →Carrying $10.3M in inventory with a fill rate of 93%. Frequent stockouts on C-category parts despite excess stock in A-category. Mathnal's Inventory Optimisation Engine recalculated safety stock using demand CV, lead time variability, and Bayesian probability. Result: 31% SS reduction ($3.2M freed), fill rate improved to 98.5%, and 47 at-risk SKUs identified that traditional methods completely missed.
Real-time Power BI dashboards for 15+ KPIs — OTIF, fill rate, inventory turns, forecast accuracy, lead time, warehouse utilisation — with SKU-level drill-down and automated threshold alerts.
Request a Demo →Replaced 23 manual Excel reports with a single live dashboard. Decision latency dropped from 12 to 3 days. OTIF improved 6 percentage points within first quarter as teams could see and act on exceptions in real time.
Autonomous AI agent that monitors exceptions, generates corrective actions, and executes pre-approved responses — from reorder triggers to supplier escalations. Your planners focus on strategy; the Copilot handles the noise.
Request a Demo →Deployed SC Copilot to monitor supply exceptions across 6,200 SKUs. 55% of exceptions auto-resolved. Planner capacity freed 40%, redirected to strategic sourcing. Decision latency dropped from 12 days to under 4 hours for critical alerts.
VRP and multi-stop route optimisation with time windows, vehicle capacities, and multi-depot scenarios. Produces actionable daily dispatch plans that reduce cost and improve delivery performance.
Request a Demo →Optimised vehicle routing across 8 distribution centres serving 2,400 delivery points. Transport cost reduced 12% ($890K annually). Vehicle utilisation improved from 68% to 87%. On-time delivery rate improved from 89% to 96%.
Continuous supplier risk monitoring with geopolitical, financial, climate, and quality risk scoring. Early warning system that detects disruptions 14+ days before they hit your supply chain.
Request a Demo →Identified financial distress in a critical single-source supplier 14 days before delivery failure. Alternative supplier activated within 72 hours. Prevented $1.8M in production downtime. Now monitors 340 suppliers across 12 risk dimensions continuously.