The Agentic B2B Sales Economy

There is a particular moment in every business transaction that we have taken for granted for millennia: the moment when two human beings size each other up. This is defined by a handshake and eye contact, the pause before the counteroffer, and the subtle dance of trust and skepticism that precedes the exchange.

We are about to lose that moment entirely.

Not because humans will stop buying and selling, but because we will no longer be in the room. Autonomous agents will act as buyer and seller, engaging in rapid, continuous negotiation via API calls, optimizing for variables beyond our understanding.

The transition from chatbots to agents is more than a technical evolution—it is a shift that redefines our economy and society. We are moving from tools that support human choices to autonomous entities making choices for us. The key question is not efficiency, but what it means for commerce to become post-human.

Efficiency as Virtue

The case for agentic commerce is seductive, and I do not dismiss it. I have been in so many negotiations in which I have dreamed of a faster, fully automated procedure, rather than listening to the buyers' monologues until 7.30pm. Travel costs for a group of four on our side include time spent at the airport preparing for the next negotiation. This has cost me a lot of time and energy, and it still costs us a lot today.

Imagine a world where procurement happens in real-time. Where your company’s autonomous agent identifies a supply chain gap at 2:47 AM, evaluates seventeen vendors, negotiates price and delivery terms, and issues a purchase order. This happens all before your head of operations pours their morning coffee. There are no delays, no cognitive bias, and there is especially no relationship fatigue.

Mechanically, this is better. Buying cycles compress from weeks to minutes. Discovery moves to data-driven APIs. Differentiation is quantifiable—uptime, defect rates, ESG scores. The agent values measurable outcomes over stories or branding.

In this vision, salespeople move from repetitive prospecting to complex deals that demand judgment and strategic thinking, while agents autonomously qualify and engage routine opportunities.

This is the vision often promoted: a frictionless, optimized, always-on economy where efficiency dominates, and the best solution wins every time—free from the perceived inefficiencies of human psychology.

What We Lose When We Automate Trust

On the surface, we are automating procurement—the mechanical act of buying goods and services. But procurement is never just procurement. Embedded within every commercial relationship is a latticework of human judgment: the decision to trust a new vendor because their account manager reminds you of your college roommate. The willingness to extend payment terms because you know the founder is going through a rough quarter. Or the choice to pay slightly more for a supplier whose factory workers are treated well, even though your CFO’s spreadsheet says otherwise.

These are not inefficiencies—they are reminders that commerce exists to serve humans, not the other way around. Human choices in commerce shape trust and connection, which automation risks erasing.

Agents optimize for what is quantifiable, but some decisions—like cultural fit or sensing future risks—defy quantification. Choosing the 'inefficient' option can build resilience and loyalty.

More troubling still, agents operate at a speed beyond human comprehension. What happens when the humans who are supposed to “set policy and guardrails” for their agents no longer understand the system they are governing? When deals are struck in milliseconds based on optimization functions that have drifted from their original intent?

Algorithmic trading shows machines can drive unintended system behaviors. Expect similar dynamics in commerce, affecting not just prices but supply chains and infrastructure.

Designing the Future of Sales Negotiation

I am not calling for a rejection of agentic commerce. The technology will be delivered with Agents who work 24/7 on B2B transactions. The question is not whether but how—and more importantly, what we choose not to automate.

Here is what that might look like in practice:

Agents for execution, humans for strategy. Let the agents handle the repetitive, high-frequency, low-stakes transactions—office supplies, cloud storage, commodity inputs. But reserve the complex, high-stakes, relationship-dependent deals for human negotiation. Not because humans are “better” at these deals, but because these are the deals where trust, context, and long-term thinking matter more than speed.

Transparency as a design principle. If an agent is making decisions on your behalf, you should be able to interrogate its reasoning in plain language. “Why did you choose Vendor B over Vendor A?” should yield an answer a non-technical executive can understand.

Intentional friction. We should establish built-in deliberate slowdowns for certain decision categories. For example, a 24-hour cooling-off period before an agent can switch suppliers, with a requirement for human review before any contract exceeding a certain threshold. These moments preserve Kairos—the qualitative, opportune moment—in a world increasingly governed by Chronos, the relentless ticking clock.

Preserve the handshake. Even in an agent-mediated world, there should be moments where humans still meet humans. Not to negotiate contracts—the agents can do that faster—but to build the deeper relationship that transcends the transactional. To ask: Are we the kind of organizations that want to do business together over the long term? Do our values align? Are we building something worth building?

This is not nostalgia. It is cognitive sovereignty, the assertion that we get to decide which parts of our lives we delegate to machines and which we insist on experiencing ourselves.

The agentic economy is being built by those who assume efficiency is the highest good and believe human judgment is a bottleneck. The main question is whether eliminating human judgment improves or diminishes commerce as a human enterprise.

But commerce is more than exchanging goods and services—it is a key way we relate as humans, build trust, signal values, and create obligations. When that relationship is fully automated, we lose more than inefficiency—we risk losing the vital human texture and meaning of doing business.

The questions to think about should be, how do you persuade a buyer that has no emotions, no ego, and no hunger?