Should-Cost Agent for COGS Intelligence
Armed buyers with real manufacturing cost data during supplier negotiations, unlocking cost reduction opportunities.
The Problem
During renewal and negotiation cycles, buyers couldn't assess supplier pricing objectively. They had no intelligence on what component manufacturing actually cost. Suppliers quoted high, buyers accepted, and millions stayed on the table. No leverage. No data-driven negotiation.
What I Built
Built a should-cost agent that ingests component specs, sourcing requirements, and market data, then reverse-engineers realistic manufacturing costs for outsourced parts. The agent analyzes material costs, labor, overhead, and supply chain factors specific to each component. Buyers input a supplier quote and get immediate intelligence on where costs should land, where there's slack, and where legitimate negotiation leverage exists. The agent pulls from multiple data sources and gives buyers a confidence-scored cost analysis they can take into negotiations.
The Impact
Shifted negotiation dynamics. Buyers walked into renewal cycles armed with cost intelligence instead of guesses. Early pilots showed 8-12% cost reduction opportunities that would have been left on the table. More importantly, negotiations became data-driven conversations instead of price haggling.
My Role
Built the agent end-to-end. Data sourcing, cost model design, agent logic, integration with sourcing tools, buyer training.
Built With
- N8N
- OpenAI
- Python