Agentic commerce: a simple announcement or a business reality?

‍A concept that drives the industry

Billed as the next e-commerce revolution, there is as much enthusiasm as skepticism about agency commerce. This approach, which promises AI agents capable of making purchases independently on behalf of consumers, is the subject of countless articles and conferences. Its visibility and place in discussions are comparable to those of a communication launching a consumer innovation at CES in Las Vegas. However, the reality on the ground tells a very different story: to date, we are only talking about potential, with very few concrete cases of relevant and viable long-term implementation of the model.

This gap calls for questions: does agency trading meet a real market need or is it a technological solution in search of a problem to solve? The lack of compelling use cases and proven business models suggests that the industry may be putting the cart before the horse.

The performance measurement frenzy

Even before considering the adoption of agency trading, we are already asking ourselves the crucial question of performance indicators. How do you measure the effectiveness of a system that does not yet exist operationally? This question reveals a paradox: the industry is questioning the KPIs of a model whose very outlines remain unclear.

This measurement problem is part of a larger challenge that digital retail is already facing: that of measuring branding in a fragmented ecosystem. Rather than speculating on hypothetical metrics for a nascent agency trade, players in the sector would benefit from focusing their efforts on solving current, very real and pressing measurement challenges.

Terminological confusion that leads to disproportionate expectations

The debate on agent trade suffers from a fundamental problem of definition. The terms LLM (Large Language Models), Generative AI, and Agentic Agents are often used interchangeably, creating confusion in the industry. This lack of conceptual clarity means that it is impossible to accurately assess the impact of these technologies on purchasing processes.

The figures concerning the influence of AI on e-commerce site traffic illustrate this confusion: the available data is not only scarce, but also contradictory. This lack of consensus on basic metrics makes any attempt to project the future performance of agency trading illusory illusory. How to measure the impact of a phenomenon whose role in the shopper experience remains undefined?

The challenge of brand visibility in LLMs

Beyond the question of agency commerce, a very real challenge is emerging: how to ensure that brands are properly represented and recommended by language models? Unlike traditional SEO, where rules and best practices have emerged over time, no structured framework exists today to optimize the presence of brands in responses generated by AI.

This lack of an optimization model for LLMs is a legitimate concern for brands and retailers. Without visibility into the recommendation mechanisms of these systems, it is impossible to build a coherent strategy. Moreover, if the evolution of ranking algorithms forced actors to regularly review their SEO strategies, what would happen for a technology that is evolving even more rapidly? The race for functional innovations of AI agents projects an unstable environment for digital marketing professionals. These questions call for answers before considering more complex scenarios such as fully automated commerce.

Focus on tangible goals

Faced with the media hype around agency commerce, the retail industry would benefit from adopting a pragmatic approach. In the end, all discussions around agency commerce overlook one central point: the shopper and its use. Before wanting to optimize a technological innovation at all costs, it is necessary to ensure that it will be adopted by its end users, that it appropriately meets their expectations and brings them truly perceived added value.

At the same time, other challenges, such as improving the digital customer experience, optimizing retail media or accurately measuring the impact of marketing investments in an omnichannel world, deserve priority attention. These issues, unlike agency commerce, have an immediate and measurable impact on business performance.

Artificial intelligence will certainly transform commerce, but this transformation will be gradual and based on concrete use cases. Before reinventing commerce, let's focus on improving what's already out there.