Companies will sell to AI customers, and human-facing websites will not be enough

On June 11, 2026, Harvard Business Review published Kartik Hosanagar’s article How Do You Market to an AI Customer?. The subtitle itself captures the change: a company must understand what technology rewards, ignores and trusts. This is a shift from traditional marketing, which optimised messages for humans, to an environment in which software takes over part of the decision. The topic follows a broader change in e-commerce. If a customer asks an AI assistant to choose a holiday, insurance, a laptop, business software or an office supplier, the seller is no longer competing only for human attention. It is competing to be included in the algorithm’s shortlist. In that world, a good-looking campaign is not enough. Data readability, clear offer parameters, availability, price, reputation and machine-comparable output matter. For marketing, this is a demanding change of discipline. Creativity remains important for humans, but structure carries more weight for an AI customer. If a product has incomplete parameters, unclear terms, poorly described variants or a mismatch between price and availability, it may be excluded before it appears to the customer. Marketing therefore moves closer to data governance, product management and operations. The business impact will be most visible in sectors where customers compare many options. Travel, financial services, electronics, insurance, software and B2B procurement all fit this pattern. An AI assistant will tend to shorten the shortlist. This may favour brands with clean data and reliable fulfilment, while poorly described products disappear from recommendations. For smaller firms, this is not only a threat. With accurate data, clear specialisation and credible evidence, they may enter the shortlist without a huge advertising budget. Legal questions will arrive quickly. If AI recommends one offer and omits another, transparency, sponsored results, possible discrimination and responsibility for bad recommendations will be debated. European rules on platforms, consumers and AI will push for explanations of why a system recommends a specific product. Companies should therefore prepare product data that can stand up not only to a customer, but also to an algorithm and possible scrutiny. Terms: An AI customer is an AI assistant that searches, compares or recommends purchases on behalf of a person. Machine-readable data are information written so that software can compare them easily. Recommendation transparency means the ability to explain why a system selected a specific offer.

Companies will sell to AI customers, and human-facing websites will not be enough

The full article is available to users registered at Hard Skills