Pricing algorithms now govern decisions once reserved for experienced executives. Machine learning models analyze competitor moves, demand elasticity, and inventory levels in milliseconds, generating recommendations that can reshape profit margins overnight. But according to Jack Truong—who led pricing transformations at 3M, Electrolux, and James Hardie Industries—most leaders fundamentally misunderstand what these systems should do.
The mistake isn't adopting AI. It's treating AI outputs as final answers rather than refined inputs for human judgment.
"Very often business leaders would just look at the results shown by AI as the 'gospel' rather than using it as one of the key information along with the business strategy to be executed in the future," Truong said. "AI should only be used as a tool that provides crystallized information to make better decisions on how to execute the business strategy to near perfection to obtain the best results."
What Gospel Thinking Costs
When Truong inherited Electrolux's North American appliance division in 2011, it was a $4.2 billion operation with declining profits. Management had accepted the narrative that North America was a mature market incapable of growth. The data supported this view—until Truong questioned whether the data reflected market reality or management assumptions.
His first meeting with global leadership established his position: "There's no such thing as a mature market, there's only mature business managers."
The distinction matters for AI deployment. If your underlying assumptions about market maturity are wrong, feeding those assumptions into increasingly sophisticated algorithms simply generates increasingly confident wrong answers. Truong didn't need better data analysis of the existing strategy. He needed to rewrite the strategy, then use data to execute it precisely.
That execution focused Electrolux on design simplicity and aesthetics rather than competing on technological features where others had advantages. Sales doubled. The company became the second-largest home appliance manufacturer in North America. None of that would have happened if Truong had accepted algorithmic recommendations built on the premise that the market couldn't grow.
The Integration Problem
AI pricing tools excel at pattern recognition. Feed them historical sales data, competitor pricing, seasonal trends, and inventory constraints, and they'll generate price recommendations optimized against those inputs. What they can't do is understand your three-year transformation plan, your positioning shift, or your deliberate decision to sacrifice short-term margin for market share in a specific segment.
"When leaders don't integrate AI information into the business strategy to formulate key execution steps," Truong said, that's when to override recommendations. "Remember that AI is just a tool."
At James Hardie, where Truong served as CEO from 2018 to 2021, he transformed a business-to-business building materials supplier into a consumer-focused brand. The company's market capitalization grew by more than $13 billion during his tenure—a 370% increase—while organic annual revenue rose 45% and net profits climbed 85%.
Those results required pricing decisions that likely contradicted what optimization algorithms would recommend in isolation. Moving upmarket to consumers meant different margin structures, different volume assumptions, different competitive dynamics. An AI system trained on the company's B2B history would have no context for evaluating prices in the new model.
The algorithm could tell Truong what price would maximize immediate revenue given historical patterns. It couldn't tell him whether that price supported or undermined the positioning shift he was executing.
When to Ignore the Machine
Truong's framework for overriding AI recommendations centers on strategic alignment. The question isn't whether the algorithm's math is correct—it usually is. The question is whether optimizing for what the algorithm measures serves the broader transformation you're attempting.
Consider product mix decisions in volatile markets. AI systems analyze which products generate the highest margins, which move fastest, which combinations maximize basket size. All useful information. But if your strategy requires establishing presence in a new category even at the cost of cannibalizing higher-margin existing products, the algorithm will resist that move every time.
During Truong's tenure at 3M, where he spent 22 years and eventually led the office products division, he faced declining Post-it Note sales as digital alternatives proliferated. An optimization algorithm focused on maximizing revenue from the existing product line would have recommended incremental improvements to paper adhesives and perhaps some defensive pricing.
Instead, Truong reimagined the product entirely. The Post-it Super Sticky line could adhere to vertical surfaces beyond paper—computer screens, walls, refrigerators. Made in vibrant colors rather than just pale yellow, the notes remained relevant even as offices digitized. The business didn't just survive; it thrived.
No pricing algorithm would have suggested that pivot. Algorithms optimize existing business models. Humans transform them.
Information Versus Direction
The distinction Truong draws between using AI as information versus gospel maps onto a broader leadership principle he applies across decision-making: "You've got to have a strategy that really takes into consideration the strengths and weaknesses of your business, and then be able to drive there," Truong explained in discussing his approach. "But then, you've got to put the execution together to get the whole company, the [whole] organization, aligned to execute toward that direction."
AI supports the execution phase. It provides "crystallized information"—patterns extracted from noise, correlations identified across massive datasets, forecasts based on sophisticated modeling. All valuable. But the direction—the strategy determining what you're trying to achieve—remains a human responsibility that requires understanding markets, competitors, customers, and capabilities in ways that transcend pattern matching.
When pricing decisions align with strategic direction, AI recommendations carry weight. When they don't, override them.
This sounds simple but proves difficult in practice. Organizations that invest heavily in AI infrastructure often feel compelled to follow its guidance to justify the investment. Executives who championed the technology adoption face pressure to demonstrate its value. Teams that spent months training models resist conclusions that their work should be overruled.
Truong's track record—including the transformation of three major corporations—suggests the discomfort is worth it. At each company, he achieved results that existing strategies and their supporting analytics couldn't have predicted. The playbook wasn't better optimization of the status quo. It was strategic repositioning followed by rigorous execution, with data serving the strategy rather than determining it.
Volatile Markets Amplify the Problem
When market conditions shift rapidly—whether from supply chain disruptions, competitor moves, regulatory changes, or demand shocks—AI pricing systems face a challenge: their training data becomes less relevant precisely when decisions matter most.
Models trained during stable periods encode assumptions about how markets behave that may not hold during disruption. Price elasticity changes when substitute products become unavailable. Historical seasonality patterns break when consumer behavior fundamentally shifts. Competitive response curves flatten when everyone's scrambling.
These are exactly the conditions where executives most want AI guidance, and where that guidance becomes least reliable without human interpretation. The algorithm can tell you what happened last time prices moved in certain ways. It can't tell you whether "last time" bears any resemblance to current conditions.
Truong's emphasis on integrating AI information into strategy execution becomes critical here. The volatile market didn't invalidate the algorithm's mathematics—it invalidated the context in which those calculations operated. Leaders who treated AI recommendations as gospel during market disruption often discovered too late that they'd optimized for the wrong objective function.
Those who treated AI as a tool—one input among many, informing but not determining decisions—maintained strategic flexibility.
What Near Perfection Requires
Truong's language about using AI "to execute the business strategy to near perfection" reveals his actual view of the technology's value. AI doesn't create strategy or determine direction. Its contribution lies in execution precision.
Once you've decided on a pricing architecture—which segments get which treatment, which products anchor your range, which channels receive what authority—AI can help implement that architecture with consistency and speed that humans can't match. It catches edge cases, maintains guardrails, adjusts tactical prices within strategic bounds, and scales decisions across thousands of SKUs.
That's not a small contribution. Execution gaps between intended strategy and actual implementation sink more initiatives than flawed strategy. But execution precision only creates value when you're executing the right strategy.
The leaders who succeed with AI in pricing aren't those who defer to algorithmic authority. They're those who establish clear strategic intent, then deploy AI to execute that intent with ruthless consistency. The machine handles the details. Humans own the direction.
Truong's insistence that AI remain a tool rather than a decision-maker reflects three decades of experience transforming businesses. Tools amplify human capability when wielded correctly. They constrain it when they determine action. The difference between the two often determines whether transformation succeeds or fails.
Media Contact
Company Name: Meridian Technology
Contact Person: Sarah Chen
Email: Send Email
Country: United States
Website: www.meridiantech.com
