#15 eCommerce Bits & Bites

I hope you are concerned, as much as I am, about the future of online buying. If not, let me introduce you to it. 😈

Imagine having a personal shopping assistant that knows your preferences, budget, and needs and it isn’t human. It is an AI-powered shopping agent. These “agentic commerce” systems can search across websites, compare features and prices, check reviews, apply discounts, and even complete purchases autonomously. This isn’t a distant future scenario – over half of consumers were anticipating using AI assistants for shopping by the end of 2025 and early data shows that shoppers who arrive via AI agents tend to be more engaged and further along in their decision.

In short, online shopping is on the cusp of automation at scale.

So my question to you is: who will we sell to when AI does the buying?

Let me dive into this further on.

For years, we’ve been obsessed with one question: How do we convince customers to buy from us? Soon, the question will be: How do we persuade machines that buy on people’s behalf?

The agents are programmed to optimise for the user’s goals – whether that’s the lowest price, fastest delivery, highest quality, or a personalised mix of factors. Unlike a human shopper who might be swayed by catchy ads or emotional brand stories, an AI agent parses facts and figures. In other words, the classic marketing playbook of catchy campaigns and brand personas might not influence an AI the way it influences a human consumer.

This raises a pressing question:

If AI agents make the purchase decisions, where does that leave brands, loyalty, and the power of persuasion?

I believe that our only solution is keeping the customer active in the process. Consumers are not passive spectators in this process – they set the parameters for their AI agents. Research and early deployments reveal that humans often encode their own brand preferences, values, and biases into the AI’s instructions.

This means brand affinity still matters, but it may manifest differently. In the past, a loyal customer would manually choose their favorite brand; in the future, they might instruct their AI to only buy that brand unless directed otherwise. Rather than being swayed by emotion in the moment, the emotion (trust, affinity, identity) is front-loaded as a parameter in the shopping criteria.

At the same time, consumers who have no strong brand attachment in a category will let the AI optimise freely, treating products as commodities. We are likely to see a split: in commodity categories (think batteries, basic household supplies), the AI’s rational criteria will dominate, potentially making brand negligible; but in high personal-value categories (fashion, electronics, health & beauty), many shoppers will still choose to have brand or experiential preferences guide the AI’s choice.

Therefore I believe we will have two types of loyalty:

1. algorithmic loyalty - a functional rather than emotional state where AI agents demonstrate persistent preference for brands that consistently optimise their multi-variable decision criteria.
2. extreme brand loyalty - Not preference. Not habit. Not “I usually buy from them.” AI agents don’t work with soft emotions — they work with rules. If a customer tells an agent “find the best option,” the brand becomes replaceable the moment another product scores higher on price, specs, or availability. The only brands that escape this optimisation trap will be the ones customers explicitly instruct their AI to choose — “buy only from this brand,” or “don’t optimize this category.” That kind of instruction doesn’t come from marketing persuasion; it comes from absolute trust built over time. This level of loyalty is a psychological relief. Customers automate decisions where they feel zero risk and zero mental effort. In that future, the strongest brands won’t win because they’re the “best option.” They’ll win because they’ve been removed from comparison altogether. Extreme loyalty is rare — and that’s the point. Only a small number of brands will ever reach this level. Just like today: only a few brands get tattooed, only a few brands become habits, only a few brands become identity. AI will magnify this imbalance. This is not unfair. It’s just more honest.

Unfortunately we get to a double-trouble as we must be able to juggle and master both. As Forter’s analysts said,

“To thrive, merchants must translate emotional brand equity into machine-readable parameters – so that both customers and their AI assistants choose them first.”

It’s a dual challenge: compete on data to satisfy the algorithm, and compete on meaning and trust to satisfy the customer’s latent desires. Interestingly, if a brand does succeed on these fronts, AI agents might actually reinforce its market position even more strongly than human shoppers would as an AI agent has no impulse to “try something new” for variety’s sake.

In essence, “winning” the AI agent is about excelling on all the practical factors that a rational decision-maker cares about, while also ensuring you’re aligned with whatever the human behind the AI cares about.

For DTC brands, e-commerce founders, and marketers, the call to action is clear: be visionary and strategic about marrying the power of AI with the power of brand! This means anticipating customer behaviour shifts, leveraging psychology and harnessing technology. The companies that win will be those that appeal to the heart, speak to the algorithm, and ultimately become the trusted choice that transcends whoever (or whatever) is doing the buying.

Further read:

• McKinsey report - The agentic commerce opportunity: How AI agents are ushering in a new era for consumers and merchants
• Forter article - Redefining customer loyalty in an AI-first world
• Futurist Bernard Marr article - Agents Are Killing Brand Loyalty And Reshaping How We Shop
• and: What Is Zero Click And Why Is It Turning Marketing On Its Head?

This newsletter is build (also) on audience request. If you have any topic suggestion or feedback, please PM me.