People often make purchasing decisions based on gut feeling – influenced by their affinity for a brand, an appealing design, or the good feeling of treating themselves to something. But what will happen when autonomous AI agents do the shopping for customers in the future? What sounds like a far-fetched vision of the future to many, is actually becoming reality. In this article, we look at what this could mean for you as a retailer and how you can prepare your business for this scenario.
From Emotional Shopping to AI-driven Transactions:
Agentic AI differs from classic, generative AI: it not only answers questions, creates content, or makes suggestions, but also acts independently – even in the real world. Consumers can simply give AI shopping agents a goal (“I need a screen that's suitable for gaming but doesn't exceed my budget”) and the AI researches, evaluates, purchases, and pays autonomously.
The first pilot projects with agentic AI are already underway: Amazon's “Buy for Me” shops with third-party vendors, Walmart's “Sparky” is growing into an autonomous shopping assistant, and Perplexity integrates the transaction directly into the search with PayPal.
When and for what purposes will consumers use AI shopping agents?
Customers are curious but cautious. According to a study by Bain & Company, many have already used AI – but autonomous purchases are still rare. Respondents said that payment security and data protection are important to them and that they are concerned about AI making the “wrong” selection of items. At the same time, clear advantages such as convenience and time and cost savings are convincing.
AI-assisted shopping is initially conceivable for products that do not pose a high risk of a bad purchase – refilling routine products such as rice, milk, or dish soap. Where style, experience, or high costs are involved – in fashion, travel, events, or luxury products – people will probably keep their wallets in their hands for the time being.
Agent-supported commerce: a game changer for retailers?
People are influenced by marketing and advertising in their purchasing decisions. But what about AI agents?
A recent study by Columbia and Yale Universities, in collaboration with MyCustomAI, investigated how different AI models make purchasing decisions in simulated online marketplaces. The result: agents are strongly guided by classic factors such as price, reviews, and placement on the page. Interestingly, the models (including GPT, Claude, Gemini) differed significantly in terms of which criteria they weighted most heavily. It was also proven that purchasing decisions could be influenced by the way retailers presented content. The researchers' conclusion: there is no universal logic according to which agentic AI selects products. For retailers, this means that optimizations can be effective. The problem, however, is that it is not transparent which AI model decides based on which factors and how content can best be optimized accordingly. Plus: AI providers could change the “rules of the game” at any time.
Opportunities and risks for retailers and brands
What does it mean for retailers when a new form of commerce emerges in which – at least in part – AI agents make the decisions, not the people behind them? As is usually the case, this development presents both opportunities and risks for you.
Risks
- Tougher price competition when agents consistently optimize for the best deal.
- Crumbling brand loyalty because autonomous AI is involved.
- Dependence on AI platforms with new, non-transparent “ranking levers.”
- Less direct customer data for retailers.
Opportunities
- High-quality traffic with clear purchase intent instead of aimless browsing.
- Higher conversion rates through hyper-personalized, AI-supported purchasing processes and well-researched product selection.
- More efficient, data-driven processes across the entire value chain.
- Better visibility for niche providers through AI optimization of content.
New strategies for retailers and brands for agent-based commerce
The challenge for retailers is to appeal to both AI-based shopping assistants and the people behind them. Find a balance between the fact-based language of AI with hard purchasing criteria and the emotional appeal to consumers that focuses on trust, loyalty, added value, and storytelling.
1. Make data & technology fit for purpose
This includes structured and consistent product data and agent-friendly APIs so that AI can access product features, prices, reviews, or availability. The payment process via AI agents must be enabled with tokens.
2. Make loyalty programs and offers agent-friendly
Discounts, bundles, and loyalty benefits should be made available via API so that agents can select them for their customers.
3. Translate values into “machine language”
Sustainability seals, quality features, or certificates should be identified as verifiable, structured signals.
4. Align monitoring with AI agents
AI search engines and LLM models are changing the KPIs that companies use to measure their success on the internet: How often is my brand cited in AI responses, or how many new users do I get through AI bots?
5. Strengthen emotional attachment to the brand
AI agents may select products based on rational criteria in the future, but the prompts will still come from humans. Strengthen your company's presence at the beginning of the customer funnel. Offer experiences and communicate your values – both online and offline, from Instagram to pop-up stores. Focus on your brand image, customer loyalty, and community building.
With all the disruptive developments of recent years, it's almost impossible to hear it anymore, but it remains true: we must see challenges as opportunities. Ravi Kumar, a machine learning specialist for the American discount chain Dollar General, develops dialogue-oriented AI models. He sums it up as follows: “Retailers need to embrace AI agent marketing – which means adapting their strategies to make their products more attractive and accessible to AI agents acting on behalf of consumers.” He expects agent-based commerce to lead to significant changes in existing business models. “However, based on my understanding and knowledge, new business models can also emerge. Agentic AI will open up new customer segments and create a level playing field for large and small retailers.”
