Position Paper: AI in Pricing

Eyes on the Stars, Feet on the Ground. Artificial intelligence is the latest cure for the common price—promises of margin uplift at the click of a button. Consultants and software firms assure executives that AI agents now out-think seasoned operators. Teddy Roosevelt had the right counsel: keep your eyes on the stars and your feet on the ground. Ambition without grounding becomes illusion, and in pricing, illusions are expensive.

I’ve been asked more than once: “So how much margin will AI give us?” That’s when I know I’m already the bad guy. Because if that’s the question, the real work of pricing hasn’t even started. AI doesn’t hand you margin. People, discipline, and judgment do.

Think of today’s AI as the most enthusiastic intern you’ve ever had. It can hoover up data, spot patterns, and draft tidy summaries. It never gets tired and always returns with something to consider. I’m glad to have that help. But I’ve never put an intern in front of a customer to run a negotiation. I don’t do it with AI either. It can’t make a salesperson enter clean CRM notes, can’t tell whether a one-off concession was a smart investment or a bad habit, and can’t read the bluff across the table. Enthusiasm without experience just makes more noise.

We’ve all heard “garbage in, garbage out.” In consumer markets—airlines, e-commerce, ride-hailing—AI thrives because the data is abundant, repeatable, and in the cloud. B2B is different. Deals are infrequent, customized, and shaped as much by relationships as by invoices. Data is scattered across ERP, CRM, and spreadsheets—much of it tribal and nowhere near the cloud when AI goes looking.

I once walked into a business where the so-called “dataset” wasn’t just scattered price lists and spreadsheets. The real inputs were previous customer disappointments, a sales leader’s fear of a competitor, over-dinner advice from an internal advocate, and emotional cries to “price to win” a so-called strategic opportunity. Historical prices and customer-sat scores lived in the cloud. The half-hearted CRM notes—“lost on price”—missed the point. More often it meant, “we failed to communicate our value.” No algorithm untangles that.

Pricing isn’t just math; it’s behavior—irrational, contextual, and contradictory. Buyers anchor to arbitrary numbers. Salespeople discount depending on confidence and incentives. Competitors move for reasons that aren’t on a supply-and-demand curve. These dynamics don’t live in datasets. They live in experience, judgment, and culture. Algorithms can recognize patterns in history, but when history is incomplete or irrelevant, pattern recognition becomes misdirection.

The bigger risk isn’t the software bill. It’s disillusionment. Executives told to expect quick margin wins from “AI pricing” often end up disappointed. Then the blame falls on “pricing” itself or on leaders labeled as skeptics. That slows improvement in one of the few disciplines that changes profits fast. The truth is simpler: AI can’t substitute for fundamentals. Without clean data, governance, and a firm grasp of behavior, AI doesn’t accelerate progress—it accelerates confusion.

This is where pricing leadership earns its keep. Inside most companies, ambition and execution drift apart. Strategy talks about better margins, faster growth, and modern commercial models. Capability is tied to messy systems, inconsistent data, and habits built over years. Pricing is the connective tissue between those worlds—the one place that translates thousands of signals, from CRM entries to customer behavior, into hard dollars. The problem is, most companies lack the discipline and accountability that make pricing stick. Drop a shiny new tool into that environment and all you’ve done is automate the dysfunction. AI doesn’t fix fundamentals; without them, it just lets you do stupid things faster.

Done well, AI is useful. It can surface outliers worth a human look, flag margin leakage earlier, and give managers a faster way to see what changed and why. But it doesn’t make the call. Leadership decides whether the pattern is real, whether it’s worth acting on with this customer, in this market, right now—and whether the action will stick. That requires roles, rules, and ownership. Who sets target prices? Who can approve exceptions? What evidence counts? Who is measured—and paid—on the outcome? AI doesn’t answer those questions. People do.

I’ve never seen an AI agent stare down a customer, pause for ten seconds, and move the price. People do that. Culture does that. That’s why AI is a tool—not the dealmaker.

So, yes: eyes on the stars. Use AI to see more, sooner. Pilot use cases where the data is ready. Treat recommendations as hypotheses to test, not orders to follow. Be explicit about guardrails, and make sure sales, finance, and product know who owns what. Keep the feet on the ground: clean up the data you actually use, tighten governance you actually enforce, and measure the dollars you actually keep.

The promise of AI in pricing is real, but only if we stay clear-eyed about its limits. AI is a powerful intern—valuable, energetic, and fast—when guided by adults in the room. Ignore the fundamentals and you’ll automate the noise. Build the foundations and you’ll get the lift without the hangover. That’s the difference between star-gazing and illusions.

Eyes on the stars, feet on the ground. Not a savior, not a threat—just a tool that succeeds in the hands of disciplined leadership. Mr. Norkus is the founder of ChiefPricingOfficer.com, a community of executives with more than 20 years’ experience leading pricing functions. He has worked in pricing consulting for over 30 years.