Blog

The Short Version

Last March, one of the most hyped buttons in e-commerce quietly disappeared. OpenAI’s ChatGPT “Buy It” feature—the one that was supposed to turn shopping into a conversation—was shut down. According to Fast Company, fewer than 30 of Shopify’s millions of merchants had ever used it.

That doesn’t mean agentic commerce is dead. Google announced its Universal Commerce Protocol with Walmart, Target, and Shopify. McKinsey says agentic commerce could be a $3 trillion to $5 trillion global market by 2030. And B2B, as usual, is rushing to “get ready” for a future where robot buyers meet robot sellers.

But ready how? And for whom?

---

Who Actually Asked for This?

The main argument for the sprint is growth. McKinsey’s 2026 research says high-growth companies are three times more likely to have increased AI investment by double digits, with benefits like higher revenue per relationship manager and 20% to 30% lower cost to serve.

Notice something? Every benefit is about seller economics. Nowhere does a buyer ask for this.

The agentic-buyer vision rests on two big assumptions:

1. Buyers need to buy so much, so fast, that buying must be automated.
   For routine replenishment, sure. That’s been happening for decades. But for considered purchases? Show me the procurement team that can’t keep up with everything the company needs to buy.
2. Sellers are so buried in deals that they need AI agents just to keep pace.
   The author has spent 35 years around B2B sales and has never met a sales team suffering from “too many deals moving too fast.” Most sales teams are trying to get buyers to move slightly faster than continental drift.

Modern B2B buying is miserable. Forrester says the average buying decision involves 13 people, and 86% of purchases stall. But why? McKinsey’s own research says the No. 1 reason buyers switch suppliers is inconsistent information across the vendor’s teams—ahead of not being able to reach someone knowledgeable.

The problem isn’t a manual checkout. It’s that 13 humans have to reach consensus while the vendor tells them 13 different stories. An AI agent can’t get a CFO to consensus. A faster pipe to a confused message just delivers confusion sooner.

And the trust math is strange: only 6% of marketers consistently accept their ad platform’s AI recommendations. Yet we’re supposed to believe buyers will let an AI agent shortlist a six-figure software vendor.

---

Marketers Are Building for More, Faster, Cheaper

If buyers didn’t ask for this sprint, who’s running it? Mostly marketers—encouraged and pressured to move fast.

The Content Marketing Institute’s 2027 B2B trends research found:

• 88% of B2B marketers say their organization uses AI-powered content creation tools.
• 60% expanded that use in the last year.
• Fewer than 4% pulled back.
• 13% say autonomous AI agents are “core to strategy,” and 46% are experimenting with them.

But when asked how AI has contributed to content operations, the answers were:

• Speed: 78%
• Efficiency: 59%
• Volume: 44%

Nobody said, “We understand our buyers better now.” Nobody said, “Our story got sharper.” Nobody said, “We finally know which half of our content works.”

The tools are doing exactly what they were sold to do: more, faster, cheaper.

---

Nobody’s Ready—Not Even the Builders

StackAdapt’s “AI Delegation Gap” study found:

• 90% of marketers use AI.
• Only half are comfortable letting it act autonomously.
• 80% feel moderate to strong pressure to increase AI usage.
• 53% say leadership’s expectations exceed what their organization is ready for.
• 6% act on their platforms’ AI recommendations “almost always.”

So we bought software with a copilot, but we don’t take its advice.

“AI agents are core to our strategy” may just be what you tell a survey when the strategy is one you’re asked to execute—not one you chose.

And the buyer? The buyer didn’t show up to the race. As Maarten Albarda points out, more than 80% of retail still happens in physical stores. Actual agentic buying is a rounding error, mostly in low-stakes replenishment. Few humans trust a bot to pick their couch.

The customer we spent a year building AI checkout infrastructure for is mostly hypothetical. Meanwhile, real human customers get less direct attention every year.

So yes: marketers are being pressured to build something almost nobody is ready for—and almost nobody really needs. It’s a full sprint to nowhere.

---

What the “(Yet)” Is For

This isn’t a rant saying agentic commerce is fake. Some version of it probably is fake. But the protocols are real. The money behind them is very real. Pretending otherwise is its own kind of hype.

The real problem is readiness in marketing and content teams—and readiness turns out to be embarrassingly unglamorous:

• Get your product data clean and complete enough for an agent to read.
• Do answer-engine work (AEO, GEO, whatever acronym you prefer) so machines summarizing your category find your brand.
• Know what your buyers actually ask—which only comes from talking to enough human customers.
• Make sure your company story is consistent enough that a machine reading your website and a prospect talking to your rep would recognize the same company.

If that sounds like classic, well-worn marketing work, that’s the point.

Here’s the weirdly hopeful part: when the robot buyer does arrive, it will be the most literal-minded customer in history. It won’t be dazzled by the espresso machine at your conference booth, flattered by a $1,000 dinner, or impressed by an orchestrated full-funnel GTM activation. It will read your data and plain-language answers, then compare them to everyone else’s.

The irony is that the best preparation for selling to a machine is the same as the best preparation for selling to a human: know your industry, your business, and your customers well enough to say true things clearly.

So don’t ignore the robot shopper. Just stop sprinting toward it with an empty wagon. Use the “(yet).” Slow down long enough to remember that we’re the ones who should choose the destination.

Speed was never our problem. Direction was.

It’s your story. Steer it well.