There’s been no shortage of talk about AI’s impact on business, but the conversation has reached a new frontier. With the rise of large language models (LLMs) like ChatGPT integrating directly with e-commerce platforms such as Shopify, products can now appear inside a chat and be purchased on the spot. It’s an impressive technical leap but it prompts a bigger, more urgent question: who decides which products you see? Writes Zane McIntyre, CEO at Commission Factory.
Since it launched to the public, we’ve trusted AI to recommend content, answer questions, and generate ideas. But once it starts recommending what to buy, the stakes shift. How does an LLM decide what to surface? Is it relevance, quality, margin potential, or something else entirely?
If this evolves into a pay-to-play model, where brands with the deepest budgets control visibility, we risk turning AI into yet another advertising channel rather than a neutral assistant. AI promised to democratise access to information. If influence is bought, not earned, we risk reinforcing the same inequalities that already shape digital marketing.
This isn’t a new concern, but it’s magnified by scale. When an LLM becomes a purchasing interface for millions, even subtle biases can distort markets. We’ve already seen how algorithms on search and social reward engagement over accuracy. If similar dynamics define AI-driven commerce, smaller brands will struggle to compete unless they can afford to train themselves into visibility.
Ironically, these models learn from us. The bias they exhibit is often a reflection of our own preferences and behaviour. But when commercial incentives enter the picture, that bias shifts from reflective to directive, shaping what consumers see and what they overlook.
For consumers, this may initially feel like convenience: ask for a product, receive a fast recommendation, buy it instantly. But if every recommendation gradually funnels toward a narrow set of brands, the diversity and discovery that make shopping enjoyable could fade.
For brands and retailers, the trade-off is more complex. AI-driven commerce promises unparalleled reach, but it also introduces a new intermediary. When a chatbot becomes the layer between brand and consumer, the story, nuance, and customer relationship risk being stripped away. The sale may remain; the experience may not.
And there are bigger risks ahead. AI could:
- Recommend competing products based on bid values or data signals, not true consumer fit.
- Present inconsistent or inaccurate representations of brand values or policies.
- Reduce differentiation, making premium or niche propositions harder to justify.
- Create dependency on platform algorithms that can change without warning.
Brands will need to choose their level of participation in this future. Will they compromise brand integrity for instant exposure? Will they allow an LLM to speak for them in moments that shape perception? As AI becomes part of the customer decision journey, transparency in how products are ranked and surfaced will be critical.
Ultimately, consumers will determine the trajectory. If they value convenience above all, AI-assisted purchasing will dominate. But if they begin to sense that recommendations are biased or bought, trust in both the technology and in the brands relying on it will erode quickly.
AI in commerce isn’t inherently good or bad. It reflects the choices we make as an industry. Transparency in how products are surfaced, clarity in how recommendations are formed, and fairness in how visibility is earned will determine whether this next phase of AI becomes a genuine step forward for consumers, or simply another layer of paid influence.
The future of AI-driven commerce is still being written. The decisions we make now will determine whether it becomes a force for transparency and innovation, or a new gatekeeper in an already crowded digital economy.
“Once AI begins recommending what to buy, it stops being a passive assistant, it becomes an economic gatekeeper.”
