Retailers moving toward AI-driven dynamic pricing often start with the wrong question. Instead of asking what technology to deploy, they should first define the pricing problem they're trying to solve, according to Gloria Garattini, Product Manager for Syntphony Pricing Management at NTT DATA.
Garattini, who has a background in mathematics and works with retailers shifting from manual pricing to data-driven decision-making, said successful pricing depends on more than algorithms. "Working across different retail sectors has shown me that technology alone is never enough. Successful pricing depends on the right balance between data, algorithm-based insights and services and business expertise."
Start with data, not models
NTT DATA's Syntphony Pricing Management supports the full pricing lifecycle, from strategy definition and business rules to demand forecasting, competitor monitoring, scenario simulation and price optimization. But Garattini stressed that AI is not a panacea - it's a tool for making better, faster and more personalized decisions in specific circumstances.
The first requirement is reliable data on transactions, product information, competitive intelligence and market dynamics. Without that foundation, even advanced models produce poor results. Companies also need a clear pricing strategy that supports commercial goals such as profitability, competitiveness, margin protection or price perception - not the other way around.
No single model works best in every scenario. Garattini recommends a multi-model approach, supported by governance and human expertise, that lets businesses select the most appropriate method for each pricing challenge. For executives weighing how to approach this, resources on AI for Executives & Strategy can help frame the decision-making process.
Leadership and metrics
Strong leadership is essential because AI-driven pricing is a business transformation, not a technology project. Pricing touches commercial, finance, marketing, category management and e-commerce teams, so leaders must establish a shared vision, clear ownership and agreed decision-making processes across those functions.
Leadership also builds trust. Teams need to understand how recommendations are generated and how they support commercial objectives. "AI should enhance pricing expertise, while strategic decisions remain guided by business priorities and human judgement," Garattini said.
Success can't be measured through a single KPI. Pricing affects profitability, competitiveness, customer perception and operational efficiency, so the appropriate metrics depend on the business objective. They may include margin performance, revenue, competitive positioning, price perception and forecast accuracy. Companies should also measure process improvements: responsiveness to market changes, consistency across channels, adoption of recommendations and decision-making speed.
What gen AI and agents add
Real-time demand forecasting helps retailers respond faster to changes in demand, market conditions and competitor activity. Gen AI makes pricing tools more accessible by letting business users explore data and understand recommendations through natural language. AI agents can continuously monitor data, detect opportunities or anomalies, simulate scenarios and recommend actions.
Garattini said the greatest value of agents today is in supporting pricing teams, not replacing them. The industry is moving toward augmented decision-making, where technology helps people make faster, more informed and more consistent pricing decisions. For operations leaders, this shift toward augmented workflows connects directly to AI for Operations.
Avoiding customer backlash
Dynamic pricing only works without backlash when it starts from a clear commercial strategy. The objective should not be to maximize the sale price of every transaction, but to balance profitability, competitiveness and customer trust.
Customers accept price changes when they are transparent, consistent and linked to understandable market conditions. Backlash tends to arise when prices feel unpredictable or unfair. AI should operate within clearly defined pricing policies, including acceptable price ranges, competitive positioning and customer-experience principles. The business sets the strategy; AI applies it consistently and at scale.
Why this matters for executives and strategy
For executives, the takeaway is that dynamic pricing success depends on decisions made before any model is deployed. Defining the commercial objective, securing data quality and establishing governance are leadership responsibilities, not technical ones. The companies that get this right will treat AI as a tool for applying a coherent pricing strategy consistently - not as a substitute for having one. The measure of success is not how often prices change, but how effectively the organization adapts to market conditions while preserving customer confidence and brand trust.
Your membership also unlocks:
