Latest / The Buyer's Side with Fexingo: Procurement, Sourcing, and B2B Purchasing Decisions / How Buyers Negotiate Against AI Price Predictions
Transcript
- Lucas: Look, if you’re still negotiating based on last quarter’s invoice or a gut feeling about where steel prices might go, you are leaving money on the table. By October 2026, the smartest procurement teams aren’t just looking at what suppliers are asking for; they are running predictive models that tell them exactly what those prices be. Luna: So you’re saying the leverage has shifted from having the best relationship with the vendor to having the most accurate data? Lucas: Exactly. It used to be about who knew the guy in sales. Now it’s about whose algorithm saw the copper shortage in Chile three weeks before the news broke. I’ve been looking at how large manufacturers are using these tools to set a hard price floor before they even sit down at the table. Luna: That sounds incredibly aggressive. Does the supplier even know they’re being out-argued by a machine? Lucas: They realize it fast enough when their opening quote gets rejected without any pushback because the buyer already knows the market rate dropped by four percent yesterday. Let’s talk about how this actually works in practice, specifically around raw materials like aluminum and lithium. Luna: Right, because those are commodities with very public, transparent global markets, so the data is there to feed the model. Lucas: Precisely. If you buy specialized software licenses or custom machinery, the data is opaque. But for bulk commodities, the AI aggregates futures contracts, shipping indices, and even weather patterns to predict spot prices. A buyer can walk into a meeting with a spreadsheet showing that the supplier’s margin is artificially inflated by twelve percent compared to the algorithmic forecast. Luna: That’s a terrifying number to hear if you’re a sales rep trying to protect your commission. Lucas: It is. And here is the thing: it’s not just about squeezing the supplier. It’s about efficiency. Procurement teams used to spend forty hours a month just gathering intel. Now, the AI does it in seconds. That frees up the negotiator to focus on terms, delivery schedules, and service levels rather than just fighting over the unit price. Luna: But what happens when the model is wrong? Because sometimes data lags behind reality, especially with geopolitical shocks. Lucas: That’s the danger zone. If the model relies heavily on historical data, it can miss a sudden regulatory change or a port strike. We saw this earlier this year when some buyers tried to lock in long-term contracts based on an AI prediction that didn’t account for a new carbon tax implementation in Europe. Luna: So the model was too smooth, but the real world got jagged. Did those buyers get stuck with bad deals? Lucas: Some did. They signed at the predicted low price, only for the market to spike immediately after. The lesson isn’t to ditch the AI, but to treat it as a baseline, not a crystal ball. You need human override capabilities built into the workflow. Luna: I’m curious about the supplier side of this. Are they also using AI to predict what the buyer will pay? Lucas: Absolutely. It’s becoming an arms race. The biggest suppliers have their own pricing engines now. So when you negotiate, you aren’t talking to a person; you’re talking to two algorithms trying to find the middle ground. It changes the dynamic completely. Luna: Does that make negotiations faster, or just more confusing? Lucas: Faster, but less human. There’s no small talk about the weather. You state your algorithmic offer, they counter with theirs, and the gap closes quickly. It’s efficient, sure, but it removes the relationship capital that used to help you get favors during tight supply periods. Luna: Which means the companies that still maintain strong personal relationships might have an edge when things go off-script. Lucas: Spot on. When the models clash or fail, the human connection is the fallback. But until then, if you don’t have ai driven pricing intelligence, you are basically guessing. Luna: Honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Lucas: Thanks, Luna. It really helps keep the research sharp. Now, back to the negotiation table. Luna: Yeah, let’s dig deeper into how buyers actually present this data without sounding arrogant. Lucas: Great question. If you walk in and say 'my computer says you should charge less,' you will alienate the supplier instantly. The best buyers frame it as shared market intelligence. They say, 'We’re seeing data from three different indices suggesting a downward trend. Can you help us align your quote with that trajectory?' Luna: That’s much smarter. It invites collaboration instead of declaring war. Lucas: Exactly. It turns the AI from a weapon into a neutral third party. You’re not attacking the supplier; you’re pointing to the data. It makes it harder for them to argue because you’re not making a subjective claim; you’re citing objective market signals. Luna: But what if the supplier has proprietary data that the buyer’s AI doesn’t see? Lucas: Then you lose leverage. This is why some buyers are starting to share anonymized demand data with suppliers in exchange for access to their production cost structures. It’s a transparency trade-off. You give them volume certainty; they give you cost visibility. Luna: That feels like a big risk for the supplier. Giving away their margins? Lucas: It is risky, but in 2026, opacity is punished. If you can’t justify your price against the market data, you get replaced by someone who can. The suppliers who embrace this data sharing are building deeper, stickier partnerships. Luna: So the winners are the ones willing to be transparent, not the ones hiding behind complex pricing formulas. Lucas: Without a doubt. Complexity is no longer a shield; it’s a liability. Buyers want clarity, and AI provides that clarity if you use it right. Luna: Do you think smaller companies can compete with this, or is this only for enterprise-level buyers? Lucas: Good point. SaaS platforms are bringing these AI pricing tools to mid-market buyers now. You don’t need a team of data scientists anymore. You just need a subscription to a good procurement platform. The barrier to entry is lower, which means more competition in negotiations. Luna: That levels the playing field significantly. Small businesses can finally stand toe to toe with the giants. Lucas: Right. And that forces all suppliers to be more competitive. It’s a healthier market overall, assuming the data is accurate. Luna: Accuracy is key. One bad data point can ruin a whole quarter. Lucas: Which brings us back to the human element. You need people who understand the data, not just trust it blindly. The job title is changing from 'Procurement Manager' to 'Commercial Data Analyst.' The skills required are shifting dramatically. Luna: So if you’re in procurement right now, you’d better start learning Python or at least understanding SQL queries. Lucas: Not necessarily Python, but you need to be comfortable interrogating your data sources. Ask where the numbers come from. Question the assumptions. Don’t just accept the output. Luna: Curiosity is still the best tool we have, even in the age of AI. Lucas: Always. Technology amplifies skill; it doesn’t replace judgment. If you have good instincts and good data, you’re unstoppable. Luna: And if you have neither, well, the AI won’t save you either. Lucas: True. But at least you’ll have accurate records of your failure. Luna: Ouch. But fair. So, how do you recommend buyers start implementing this if they’re currently running everything on spreadsheets? Lucas: Start small. Pick one high-volume category, like packaging or logistics. Integrate a pricing tool there first. Prove the value, then expand. Don’t try to boil the ocean. Luna: Practical advice. Build the muscle before you lift the heavy weight. Lucas: Exactly. Get comfortable with the insights, then scale the confidence. It’s a journey, not a flip of a switch. Luna: I love that. Slow and steady wins the race, even when the race is digital. Lucas: Well said. As we wrap up, think about your next negotiation. Do you have the data to back up your position, or are you just hoping for the best? Luna: That’s a great question to leave with. Maybe we should all check our models before we check our calendars. Lucas: Or maybe just call your supplier and ask how their day is going. Sometimes the old ways still work. Luna: Don’t tell the AI that. Lucas: Too late. It probably already knows.