Data Is Cheap. Judgement Is Expensive.
Every procurement team I talk to is drowning in data. Commodity price feeds, weather forecasts, freight indices, supplier financials, tariff updates, news alerts—you name it. Most of it costs next to nothing to collect, and a lot of it lands in your inbox every morning whether you asked for it or not.
Yet the hardest part of the job hasn't changed one bit. Someone still has to look at all these numbers and make a call. Do we buy now or wait? Should we cover 40% or 70% of our demand? Do we push back on the supplier or take the price hit? Do we switch origins or stay put?
Data is cheap. Judgement is expensive. I learned that lesson long before anyone started talking about "AI in procurement," and it's the exact reason we built Kodiact.
What a trading desk teaches you
I started out as a category manager trading commodities for a food manufacturer, and I learned quickly: two people can look at the exact same screen, the same market prices, and the same reports, yet come to completely different conclusions. One person makes money for the company; the other loses it.
The difference was never about who had better data. Everyone on the desk had access to the exact same feeds. It always came down to real-world judgement:
- Knowing which signal actually matters today and which is just daily noise.
- Understanding what expectations the market has already factored in.
- Knowing your own operational reality—plant schedules, existing inventory, open contracts, and budget targets.
- Sensing when it's time to move and when it's better to hold off.
The best traders I worked with had lived through a few market cycles. They remembered how a drought in a key region played out years ago, how suppliers acted when prices plummeted, and how long a price spike took to filter into finished goods. They held all this valuable context in their heads—because nobody ever wrote it down.
What direct materials procurement teaches you
When I moved from the trading desk into broader direct materials procurement, the scope grew significantly. A single decision about a key ingredient now impacted supply security, vendor choice, product specs, warehouse inventory, contract language, and the bottom line for dozens of SKUs. The judgement required went far beyond tracking unit prices.
I saw three major pain points keep popping up.
First, category strategies went stale fast. Teams would write a strategy document once a year, present it to leadership, get approval, and put it on a shelf. By Q2, market conditions had shifted entirely, but the document stayed frozen in time.
Second, critical knowledge was locked in individual heads. A veteran category manager knew exactly why the company dual-sourced a critical component, or why the team never let coverage dip below a certain threshold before harvest season. When that person retired or moved on, all that institutional wisdom walked out the door with them, leaving the next person to start from square one.
Third, decisions weren't captured. Companies were diligent about saving final contract prices and POs, but almost nobody logged why a decision was made, what options were considered, or whether the outcome matched expectations. Without that context, teams had no structured way to learn from their past choices.
Why technology is finally stepping up
I spent the next 20 years in enterprise software, mostly building and scaling AI platforms. At Seal Software and Evisort, we used machine learning to read complex contracts and pull out actionable data. That experience showed me what AI does brilliantly—reading, connecting, sorting, and analyzing vast amounts of information far faster than any team could. But it also showed me where AI needs human direction: weighing trade-offs, determining priorities, and making the final call.
For decades, procurement tech focused almost entirely on the cheap part: gathering more data, building more dashboards, generating more reports. The expensive part—human judgement—remained scattered across spreadsheets, email threads, and informal conversations.
Now, technology has finally caught up to the real challenge. AI can handle the heavy lifting of pulling context together and modeling scenarios, freeing up experienced professionals to focus on making smart calls. And because modern systems can record those decisions, teams actually get smarter over time instead of repeatedly relearning the same lessons.
That's precisely why we created Kodiact—a decision intelligence platform tailored for F&B and CPG manufacturers buying raw materials, ingredients, and packaging. We grounded Kodiact in three core principles.
1. Comprehensive context aggregation
Good decisions require the full story. Kodiact brings together your internal enterprise data—spend, contracts, bills of materials, technical specs, and current inventory—and connects it with over 160 external feeds covering commodity markets, weather patterns, freight rates, energy prices, trade policies, and supplier risk factors.
It's all about making connections. A distant drought, a new tariff, or a refinery shutdown doesn't tell you much on its own. But when you link that event directly to your supplier origins and bills of materials, you immediately see which SKUs, production plants, and contracts are exposed—and by how much.
2. Live category strategy and real-time simulation
Annual strategy documents belong to a bygone era. In Kodiact, your strategy is dynamic and updates in real time. When market dynamics shift, the platform updates your strategy options so your team can see what changed and why.
Scenario modeling lets your team test potential moves before committing resources. What happens to cocoa costs if El Niño hits West African crop yields? How does a 10% resin increase impact packaging costs across specific facilities? What if we shift 20% of volume to a secondary supplier? Your team gets clear dollar impacts alongside actionable mitigation strategies long before a supplier sends a price increase notice.
This is where that trading desk mindset really shines. The best traders ran mental simulations constantly. Kodiact runs them instantly for your entire team using full context.
3. The decision ledger
The decision ledger is the feature I wish I had early in my career.
Every key decision gets captured: the market triggers, the options evaluated, the data available at the time, who made the call, the underlying rationale, and the eventual outcome. Over time, this ledger transforms into your team's living institutional memory.
- This fundamentally changes three key areas:
- Knowledge sharing becomes natural. A new category manager can review how experienced colleagues handled a similar scenario in the past and see why they made that call.
- Decision-making continually improves. Teams can compare projected outcomes against actual results to evaluate what worked, identifying clear patterns across categories over time.
- Institutional knowledge stays in-house. When team members move on, their strategic reasoning remains accessible to the business.
The ledger continuously sharpens the system, too. Every logged decision helps refine future recommendations, creating a virtuous feedback loop: richer context, sound reasoning, human insight, and compounding institutional knowledge.
Empowering human judgement
In Kodiact, category teams retain full decision rights and strategic control. The software manages the data gathering, cross-referencing, and complex modeling, while people make the decisions, own the outcomes, and articulate the strategy.
A skilled trader with better market context makes smarter trades. A great procurement team equipped with live strategy, real-time simulation, and a shared history of past decisions acts faster and avoids repeating past mistakes. Technology lowers the friction of applying good judgement, making human expertise far more valuable.
The real competitive edge
Every competitor in your industry can subscribe to the same market price feeds. Every supplier now arrives at negotiations armed with sophisticated pricing models. Raw data alone won't set top-performing procurement teams apart from the rest.
Judgement will—specifically, expert judgement applied quickly, shared across the organization, and refined with every single decision.
If you manage a direct materials team and this resonates with your daily reality, I'd love to connect and start a conversation.