There’s a lot of noise right now around phrases like “AI-native consumers” and “agentic commerce.” At first glance, it all sounds intriguing. Then, fairly quickly, you realize it’s mostly a familiar story with new branding.
Everything has changed… kind of
Advertising has been around for a long time. Sometimes we benefit from it, but mainly it's irrelevant and annoying. When it’s useful, it can genuinely help us discover something we want or need. That being said, the idea that the shift from traditional advertising to AI-mediated advertising and “digital discovery” is somehow exceptional feels like a well-executed ad campaign for AI and advertising itself. What’s striking is how often this new advertising model is presented as both transformative and too complex for consumers to meaningfully understand. We’re told not to worry about the details, and to just trust that the AI will find the best product for us. In reality, most of the complexity lives on the brand, platform, and ad-tech side and it's novel for us but changes little in terms of what we buy.
None of what we are being promised is really new. Personalized recommendation systems in the mid-to-late 1990s were framed as revolutionary (and to some extent they were, but they weren't and usually aren't very accurate). The influencer economy a decade ago was supposed to fundamentally change how people bought things. In practice, the consumer experience barely shifted. People still bought stuff. Discovery moved around. A lot of good influencers got paid. The real change happened in the backend, not at the front. We, the consumers, bought stuff, it was just from a live ad or recent post.
Now though (again) we are being told how lucky we are that ads are evolving, that chatbots will now understand us and serve better product recommendations. Maybe it is interesting that we can chat conversationally about something we might want to buy with a system that is, increasingly, financially incentivized to sell us something. Companies are already paying for placement; the incentives are not subtle or hard to follow and only the top performing companies can afford to compete. Ergo: we will buy stuff from the same places we currently buy stuff (at least for the foreseeable). Talking to a chatbot is marginally better than skimming reviews or scrolling through obviously-sponsored posts, and it may be genuinely useful some of the time. The actual "new thing" to be excited about it that we will soon receive truly personalized ads. But limitations haven’t gone away. The ads will just be delivered in a more personable voice.AI still hallucinates. It still panders. Reinforcement mechanisms still encourage agreement and affirmation (read: hello, echo chamber).
Personalized Ads
A note: truly personalized ads can't really be borne from ChatGPT when it's Meta, Amazon, Google and Reddit combined that know most about you (insert your own orgs and corps above for a more accurate account). And why would any of these share your data with OpenAI or any other AI co? Then they'd lose their home advantage. So we should expect those companies to run ads for us on their platforms. Realistically, nothing knows you better than your devices so an agent that quietly runs calls to recommend things for you to buy, see and do might be how this all shows up. Which brings us to "agentic commerse"
Agentic Commerse
Agentic commerce is, apparently, the next phase of digital shopping where AI agents act autonomously on behalf of users to search, compare, negotiate, and complete purchases. We might get there, but what likely comes before widespread “agentic commerce” is simply companies outspending other companies. High CPMs mean only brands with sufficient budgets will show up in these environments. OpenAI is reportedly asking $60 per CPM, which is double the current average of $30). Chatbots may synthesize options for you inside a dialogue, but the competitive dynamics for brands don’t magically improve and the underlying consumer needs don’t change either.
That being said, if platforms fully replace browsers, that might alter what we click on (or whether we click at all). Google has been circling this idea for years, though it would require leaving parts of its legacy business (pay-per-click Ads via the browser) behind, which it is already slowly departing from. Alternatively, agents might get bolted into browsers or (as already alluded to) embedded directly into phones - Apple and Google’s latest attempt to own the interface layer.
Conversion rates may rise, which is great for brands. Eventually, your personal AI agent may choose a retailer on your behalf based on some opaque notion of “fit.” Or it may reflect the incentives of the platform hosting it. Butt none of that meaningfully changes what you buy.
Agentic commerce might ensure you never run out of Uncrustables, milk, or toilet paper. It optimizes logistics and replenishment. It might change how you buy things, but not really why, and not really what.
As usage concentrates around a small number of platforms, the business model pressures are predictable. Consumer-scale AI infrastructure is expensive. Subscriptions and enterprise licensing alone can't sustain it. Just as with search and social, advertising is emerging as the monetization layer—quietly shifting LLMs from ostensibly neutral copilots into commercial systems.
Do I sound cranky? I don't know if I am, but this does feel like I'm being pre-sold something, told it's for my benefit and that it will change everything for me. I'll still have the same money in the bank and (more or less) the same things I want to impulse-buy and the same things I need each week. The way I am being sold it makes me think it justifies the enormous sums companies are paying and being asked to pay to advertise.
.jpg)
