Deepflow AI Demand and Raw Material Price Forecasting

Impactive AI Co., Ltd. was established in July 2021 in Korea, with offices in Gangnam, Seoul and at the Handong University Business Incubation Center in Pohang; its CEO is Doohee Chung. It helps manufacturers, food and beverage, retail, steel, construction and EV battery companies that rely on intuition or ERP reports for ordering, which leads to stockouts, overstock and poor timing of raw material purchases.

Deepflow is a SaaS forecasting suite built on deep learning and transformer time-series models, backed by 17 granted patents. Deepflow Forecast predicts item-level sales and shipments 6 to 12 months ahead with inventory dashboards, driver analysis and S&OP plans; Deepflow Materials forecasts prices of steel, non-ferrous metals, energy and agricultural commodities; NPD forecasts new products and MarketTwin (beta) simulates one million AI consumers. Clients include Ildong Foodis and Ecopro Materials.

• Demand forecasting: item-level sales and shipment forecasts for the next 6-12 months.
• Inventory management: dashboards flagging overstock and stockout items.
• Forecast driver analysis: explains drivers and risk factors and recommends next steps.
• S&OP support: sales and production plans built on AI forecasts.
• Deepflow Materials: raw material price forecasts with indicator analysis and buying signals.
• NPD and MarketTwin: new product forecasts and simulation with one million AI consumers.

• A Vietnamese steel or construction firm can time purchases of rebar, coil and coking coal with price forecasts.
• A food or beverage producer can forecast SKU demand 6-12 months ahead to plan production and packaging.
• A pharmaceutical distributor can reduce stockouts and expiry losses.

• A steel manufacturer cut stockouts by 35.4% and overstock by 32.2%, saving 21 billion KRW in inventory costs in November 2024 alone.
• A pharmaceutical company reduced stockouts by 22.6% and overstock by 32.5%.
• Ildong Foodis reduced inventory risk by 26%; a manufacturer cut order planning time to one-fifth.
• A construction company achieved over 96% average accuracy forecasting rebar, copper, hot-rolled coil and coking coal prices.

Raw material price volatility and demand swings push Asian manufacturers toward AI forecasting delivered as affordable SaaS rather than large planning suites. Steel, construction, food and battery supply chains in Southeast Asia, including Vietnam, are exposed to the same commodity price risks that Deepflow Materials targets.

Established July 2021; CEO Doohee Chung. Series A of 8.2 billion KRW in August 2025; selected for Korea's Super Gap Startup 1000+ and KOICA CTS Seed 2 programs; Ministry of Science and ICT Minister's Award 2025. Holds 17 granted patents from 71 applications, including 7 US applications. Partnership with PGS in Singapore in August 2026.