Kodiact launches an AI-native simulation platform for direct materials: see the shock before it hits the P&L
Chicago startup Kodiact launched an AI-native platform on September 30 that simulates price shocks across direct materials — the ingredients and packaging that can account for up to 60% of a manufacturer's revenue — weeks before they reach the P&L. The pitch: procurement doesn't need another dashboard; it needs a “system of judgment.”
On September 30, Chicago-based Kodiact announced the launch of what it calls the first intelligent simulation platform for direct materials — an AI-native product built initially for food, beverage and CPG manufacturers, but designed for any industry where raw inputs drive the business. The platform connects every material to its underlying price drivers and every decision to a unified ledger, with the stated goal of showing operators what will happen next, not just what has already occurred.
The problem it targets is expensive and measurable. Kodiact says manufacturers can buy the same ingredient from the same supplier in the same month and still pay a 16% different price — and 62% of that gap persists a year later, driven by upstream volatility that traditional category management tools fail to catch. Direct materials can represent up to 60% of manufacturing revenue, yet procurement still runs on backward-looking systems: ERPs show past payments, spend cubes re-slice the same history, and dashboards visualize it in color.
Excel still runs the category desk
The numbers Kodiact leans on come from The Hackett Group's 2026 Procurement Key Issues Report: 69% of organizations run category management on Excel, only 9% use a dedicated category management solution, and just 6% reach "Leading" maturity. Meanwhile procurement workload is projected to rise 8.0% while staffing falls 0.9% — widening a productivity gap at the exact moment supply continuity is the top priority.
"Kodiact exists to solve the hardest problem in procurement: direct materials volatility, complexity and risk," said co-founder and CEO Steve Tucker. "Manufacturers have been forced to make multimillion-dollar decisions with fragmented data, static strategies, and no way to see the ripple effects."
Three layers: what is true, what you do, what happens next
The platform is built on a single canonical model with three layers. The Data Layer holds a continuously updated representation of every material, input cost chain, supplier position and margin floor, replacing fragmented spreadsheets. The Operating Layer turns strategy and decisions into what the company calls a living system: auto-updating category strategies, automated project analysis via agents, and a complete decision ledger. The Simulation Layer is the differentiator — a simulation engine that models multi-variable shocks across materials, suppliers, demand and energy, so teams can watch ripples develop weeks before the P&L is impacted.
On the enterprise-AI question of the moment — where your data goes — CTO and co-founder Jonathan Lee made an explicit pledge: "Unlike standard LLMs, Kodiact guarantees strict data isolation, ensuring customer data, strategies, and supplier information are never mixed or exposed to shared public models."
Early pilots, and a Hackett nod
Kodiact says the platform is already in use across multiple FMCG and CPG manufacturing pilots, where teams are reducing category cycle times from 8–12 weeks to minutes, automating strategy generation, and uncovering hidden supply risks. Those are the company's own claims from the announcement — no named customers or pricing were disclosed. The Hackett Group has separately spotlighted the approach; senior research director Bertrand Maltaverne is quoted in the release: "As the saying goes, vision without execution is hallucination. Kodiact finally brings a solution to direct materials that goes far beyond just category management."
The roadmap adds operating modules — Marketplace, Trade Desk, Planning and Innovate — over the next 12 months. The founders come with domain pedigree: Tucker began his career in commodity procurement at Mars and has spent three decades in enterprise software, according to an interview with Expana.
The bigger picture: procurement AI is shifting from dashboards that describe the past to systems that simulate the future and keep a ledger of why each decision was made. If Kodiact's simulation engine works as advertised for volatile ingredient categories, the 69%-on-Excel figure becomes the market to take. The proof will be in named deployments and audited savings — neither of which arrived with the launch.