Pricing used to be a decision. Increasingly, it is becoming a prediction.

AI-powered systems can analyse demand, stock levels, competitor activity, purchase history and customer behaviour, then recommend or automatically apply a price. That makes pricing faster, more responsive and potentially more profitable.

It also creates a serious business risk.

Customers may accept that an airline ticket costs more during a busy weekend. They may feel very differently if they suspect a retailer has charged them more because an algorithm believes they are willing to pay it.

The technology is moving quickly. The rules, customer expectations and internal controls around it are struggling to keep up.

Dynamic pricing and personalised pricing are not the same

These terms are often used interchangeably, but the distinction matters.

Dynamic pricing changes according to market conditions. A business might lower a price to clear excess stock or increase it when supply is limited and demand is high. The price is responding to the situation around the product.

Personalised pricing changes according to information about the customer. That could include location, browsing history, previous purchases, inferred income or how likely someone appears to be to compare alternatives. The price is responding to the person.

AI can support both models. It can find patterns that would be difficult for a human team to identify and estimate how a price change might affect demand, revenue or margin.

But once personal information enters the calculation, the commercial and reputational stakes change.

Pricing is becoming a customer-data decision

In January 2025, the US Federal Trade Commission published initial findings from a study of companies that provide pricing technology. It found that information including precise location, browsing history, shopping behaviour and even mouse movements could be used to tailor prices and promotions. The intermediaries examined by the FTC worked with at least 250 clients across sectors including grocery, apparel, financial services and travel. 

This is bigger than a pricing-team issue. It connects customer data, ecommerce, marketing, legal compliance and brand strategy.

A promotional platform might describe its output as an optimised offer. A customer who discovers that another person received a better price may call it unfair.

That gap in perspective is where the risk sits.

Regulators are paying closer attention

On 31 August 2026, the FTC announced that it was seeking comments on a proposed enforcement policy for personalized pricing. The proposal says that businesses may risk violating consumer-protection law if they use personal data to determine prices without clearly explaining that the price is personalised, why it is personalised and what types of data are being used. 

The proposed policy does not say that all personalised pricing is prohibited. It focuses on whether customers are misled or treated unfairly.

That should be a warning against waiting for a final rule before acting. Regulators may differ between markets, but the underlying expectation is straightforward: if your pricing system uses customer data in a way people would not reasonably expect, hiding that fact is a dangerous strategy.

Businesses should obtain advice for the countries and sectors in which they operate. More importantly, they should decide what they believe is fair before a regulator, customer complaint or viral post decides for them.

The marketing team needs a seat at the table

AI pricing may be purchased by an operations, revenue or ecommerce team. Its effects will be felt by marketing.

Pricing communicates what a brand stands for. It influences customer acquisition, conversion, loyalty and word of mouth. It also needs to remain consistent with advertisements, email offers, landing pages and sales messages.

Imagine promoting an accessible fixed price, only for the checkout system to show different customers different amounts. Even if the variation has a technical explanation, the experience can make the campaign feel misleading.

Marketing leaders should know:

• Whether prices vary by customer, location, channel or time
• Which customer information contributes to those decisions
• Whether advertised prices always match checkout prices
• How loyalty offers are separated from individual price optimisation
• What explanation customer-service teams can give
• Who can pause or override the system

If those questions cannot be answered clearly, the business is not ready to automate the decision.

Do not let the software set your principles

A vendor can optimise for the objective it is given. It cannot decide what your brand should consider acceptable.

Before deploying AI pricing, create a written policy covering the data and tactics that are allowed. For example, adjusting prices because stock is approaching an expiry date may be useful to customers and the business. Increasing someone’s price because their behaviour suggests an urgent personal need is much harder to defend.

Set boundaries around sensitive information, customer vulnerability, maximum price movements and the frequency of changes. Establish which decisions require human approval.

Ask vendors exactly what data their product collects, purchases or infers. Find out whether your organisation can inspect the factors behind a recommendation and whether the platform records every price change. A polished dashboard is not a substitute for accountability.

Measure customer impact, not just margin

An AI pricing pilot may look successful if average order value rises. That is only one part of the result.

A stronger measurement plan should also examine conversion, repeat purchases, complaints, refunds, abandoned baskets and customer-service contacts. Compare outcomes across locations and customer groups. Look for sharp differences that cannot be explained by legitimate commercial factors.

Keep a record of the price shown, when it was shown, the data or rules behind it and whether a person approved the change. If a customer challenges a price, the business should be able to investigate without reconstructing the decision from memory.

It is also worth testing communication before launch. Can customers understand why prices vary? Is the explanation visible before they make a decision? Does it still sound reasonable when written in plain language?

If the explanation only works when buried in a long privacy policy, that is useful feedback.

Smart pricing needs strong guardrails

Recent reporting shows that customer concern about AI-assisted pricing is rising as digital shelf labels, algorithmic recommendations and personalized offers become more visible. The same reporting also makes an important distinction: digital pricing technology does not automatically mean that every shopper receives an individual price. Suspicion grows when businesses fail to explain what the technology actually does. 

That gives responsible businesses an opportunity.

AI can help companies understand demand, manage inventory and make better-informed commercial decisions. But the winning approach will not be to squeeze the maximum possible amount from every transaction. It will be to improve pricing while protecting the relationship that makes future transactions possible.

The question is not simply whether AI can change your prices in real time.

It is whether your business can explain, monitor and defend the way it does it.

If AI is changing how your business reaches, converts or communicates with customers, The GO! Agency can help you turn that change into a clear, credible marketing strategy. Get in touch to make sure the experience customers see matches the business you want to build.