Agentic AI is an amazing field of work, and its materialization in Agentic Commerce is even more breathtaking. Saving time and offloading all the unsatisfying tasks to an AI, whatever that really means, is the kind of promise that blinds us and prevents any hard look at what it truly implies. Because we all, as consumers and users, want it to happen.
And of course, dream sellers and fear‑mongers jump on this desire to make business out of it, throwing the whole market into a whirlwind of false promises and performative enthusiasm, all to avoid missing out on “the next big opportunity.”
That’s why we’re all talking about AI morning to night, when at a macro level there is still so little we can actually do with it. Sure, we can code faster (supposedly “at light speed”). But does that really change anything? Did faster coding suddenly give us flying cars?
As the author of The Hard Thing About Hard Things would say, the truth sits in the details. Which is also why I end up writing these borderline‑condescending posts on LinkedIn: to make people pause, breathe for two seconds, and reflect on where we actually are with AI today.
So let’s take a simple example of Agentic Commerce tackling one of the most critical yet least satisfying tasks of everyday life: repeat grocery shopping.
I’m not talking about that one special event where you put love and creativity into the kitchen. I’m talking about the weekly grind: the boring, time‑eating, utterly uninteresting but still somehow vital routine.
Here’s how the “magic” supposedly works: you have your weekly list, you confirm it, and you head to the grocery store to serve your sentence. Now you dream of simply talking your way out of it with an agent that does the whole thing online and gets everything delivered to your doorstep. That’s the target.
But the process is not straightforward. It has many crossroads, and that’s where the hard truth kicks in.
Once the so‑called Agentic Commerce “AI” (throwing in more AI is always trendy, and it makes it sound almost human, which is another story entirely) is fed with your grocery list, the fun begins:
- You need to provide product alternatives because your favorite items aren’t always in stock.
- One alternative might not be enough, or maybe you want to give the “AI” (let’s call it a person named AI, since that’s what everyone does) the freedom to decide for you. Which is how your pasta might become toothpaste — the beauty of Generative AI, apparently.
- Once you’ve survived that, you have to define a price range for every single item.
- And for the alternatives too. More decisions.
- If the agent is smart enough, it will then ask you about pack architecture: if there’s no 4‑pack yogurt, can it buy a 6‑pack? Should it look for value packs? Always or only sometimes?
- Then you get to run the whole thing again — including price ranges.
- And then you add boundaries so you don’t drown in the complexity (you already kind of are).
- Once you’re done with all that input, you still need to deal with delivery slots and payment.
- Delivery slot is easy. Payment is… something else entirely.
Payment looks simple, but in real life it’s long, painful, and full of edge cases. Can the AI pay for you in all situations? Should it check your balance? Should it pick the card on its own? And what happens when the delivery isn’t what you expected?
Who do you complain to? Yourself, because the agent was “yours”? The agentic commerce provider? Who insures this mess?
All that to say: we are still very far from full Agentic Commerce.
Not because AI is too weak. But because the real challenge lies in business process design and in interacting with companies that are simply not ready. I repeat: companies are not ready.
And I’m not talking about the design of their websites. I’m talking about their data interfaces. Agentic Commerce is the disintermediation of the website itself, and once you remove the pretty front end, you expose the chaos underneath: messy data, inconsistent structures, and systems that can barely talk to each other, let alone to an autonomous agent.
So before anything else: let’s do our homework and stop pretending.
The vision, the ambition, and the targets are clear.
Now let’s actually dive in.
Leave a Reply