For a home-improvement distributor, pricing was a war of nerves. When Amazon or a regional competitor dropped the price of a power tool by $10, it would take the client's team two days to respond. By then, the sales window was gone.
Responding to every price drop is a "race to the bottom" that kills margins. The challenge was identifying which price drops needed matching and which were just noise based on the client's current stock levels and shipping costs.
01. Competitive Intelligence Mesh: We deployed a fleet of distributed scrapers that monitored 20 competitors every 5 minutes. This data was fed into an Elasticsearch cluster designed for sub-second lookup of pricing trajectories.
02. Dynamic Elasticity Model: Using CatBoost, we trained a model on historical sales sensitivity. The system now knows if a $2 drop in price will actually drive volume, or if the customer at that price point is insensitive to small shifts.
03. The Pricing "Kill Switch": We implemented an automated governance layer. If the AI suggests a price below a certain 'Floor Margin' (calculated dynamically by shipping/COGS), it triggers an immediate exception for human review instead of blindly following the market down.
The client now responds to market shifts in under a minute. More importantly, they avoid matching price drops on items where they have low inventory or high shipping costs, effectively protecting over 7% of their bottom line.
← Return to Intelligence Engine