The Wheel of Persuasion in the Age of Agents

People do not buy your product. They buy the outcome of your product. An agent changes nothing about that.

I work at the intersection of product, GTM and AI, and I often hear the worry: now that AI agents do the first touch, the human part is lost. Persuasion turns into a machine. That is wrong. The psychology of persuasion stays the same. What changes is only who executes which part, and at what scale.

A good frame for this is Alex Cattoni's Wheel of Persuasion. Six steps that show how you guide someone from problem to product. I look at it as an operator and ask at each step: does the agent do this, or the human?

What you'll learn

• The six steps of the Wheel of Persuasion.
• Which steps an AI agent takes over and which the human keeps.
• Why the psychology stays even as the execution scales.

The thesis: persuasion follows six steps. Agents scale the early ones, humans close the late ones. The order stays the same.

🧨 The problem: companies talk about themselves

The most common mistake in B2B: companies talk about themselves, not the customer. The website is full of we, us, our. Why the company solves the customer's problem stays unclear.

There is a simple test, the we-test. Go to your website and count how often we, us, our or the company name shows up, against how often you address the customer directly. If you talk about yourself more than half the time, you have a problem. The customer does not feel picked up and clicks to a competitor.

The Wheel of Persuasion flips that. It starts with the customer and only reaches the product at the end.

This mistake is not new. The scale is. With agents you can now roll out talking about yourself a thousand times over. Or you use the same scale for the opposite. That is why it pays to study the craft of persuasion before you automate anything.

🛠️ The six steps, split up

Cattoni's wheel has six steps. I will run through them and say, for each, who executes it in 2026.

• 1. Problem. Every solution starts with the problem. Pick the customer up where the pain arises. "I help target group with problem." An agent can scale the opening, cleanly researched.
• 2. Pain. Name the negative consequences of the problem, concrete, no platitudes. The customer should feel that you know their exact pain. An agent can prepare this too, if it has the context.
• 3. Prescription. This is not your product yet. It is the method, the new approach that contributes to the solution. "To overcome your problem you need your unique solution."
• 4. Pivot. Now you shift the focus from the customer to you. Who are you, why are you the right one to talk to? This is where it gets personal, this is where the human comes in.
• 5. Positioning. Set yourself apart. Why have other solutions not delivered so far? "Maybe you have tried an alternative solution, but ran into another problem."
• 6. Product. Only now the product. Show why it solves better, easier or faster. This is often the moment a human closes.

Watch how the wheel turns: from the customer's problem through the method to you and your product. The first steps are research and outreach. The last are trust and the close.

🩺 Diagnosis first, then the prescription

Step 3 of the wheel is called prescription, and that is exactly where most B2B teams tip over. They prescribe before they have examined. Jacco van der Kooij puts it in one line in The SaaS Sales Method, his book at Winning by Design: prescription before diagnosis is malpractice.

At the doctor's you would get up and leave. In sales it happens every day. The first email pitches the product. The first demo shows every feature. And nobody asked where it hurts. The wheel forbids exactly that: problem and pain come before prescription. For demos, van der Kooij recommends 3 demo points in 30 minutes, 5 in 60. More is prescription without diagnosis.

He carries the thought through to the close. For him, 7 moments really matter in the buying process: outreach, first contact, discovery, close, kick-off, review, renewal. The trick: he phrases each moment from the buyer's perspective. Discovery then does not mean "we qualify you". It means "help me understand what I have".

It is clearest at the close. From the buyer's perspective the close means: "Help me sell this internally." Your champion, the person fighting for your case internally, has to repeat your story in front of the CFO, IT and procurement, without you in the room. If your pitch only works in your mouth, this is exactly where you lose.

For the agent question this means: an agent prepares the diagnosis. Collect signals, research context, draft hypotheses about the pain. One signal is a guess, five are a pattern. The diagnosis itself is made by the human in the conversation. And the material your champion sells with internally is something you build deliberately: one page, their language, their numbers. The psychology stays, the execution scales.

🗣️ TALKER and ACE: the blueprints for the conversation

The wheel delivers the dramaturgy. But what do you do in the minute a human actually picks up? Van der Kooij has blueprints for that too. They are the operational layer of persuasion, and they are deliberately mundane.

For the opening: ACE. Appreciation, thank them for the time. Check time, confirm the slot still works. End goal, say what should stand at the end of the call. 3 sentences, and both sides know where they are.

For the conversation itself: TALKER. Tone, the tone builds the relationship. Ask questions, ask instead of pitching. Listen actively, listen instead of waiting for your cue. Keep notes, capture what was said. Elaborate, go deeper on what the customer touches. Repeat, play it back in your own words until the customer nods. Van der Kooij's point behind it: a sales conversation is a conversation, not a qualification interrogation.

Now the agent lens. Keep notes is agent work today: transcript, summary, CRM entry, follow-up draft. An agent can also prepare good questions from the account context. Tone and listen actively no agent can take off your hands. So the blueprints confirm the thesis of this article: the machine scales preparation and follow-up. The human runs the conversation.

❓ What persuades in the end: emotion or reason?

Both, but in different places. Van der Kooij distinguishes emotional impact, which first benefits a person, and rational impact, which first benefits a company. Persuasion needs both. The person has to want it, the company has to be able to run the numbers.

The rational impact is the business case: revenue up, costs down, risk smaller. Without it no deal gets through procurement. The emotional impact is the reason your contact starts the case at all. Less stress. A visible win. A team that stops cursing.

The wheel already contains this split. Pain is emotional, a person feels it there. Positioning and product are rational, a company calculates there. Argue only rationally and you bore people. Argue only emotionally and you fail at the CFO, at the latest.

The division of labour applies here too. Rational impact is agent terrain: collect numbers, build comparisons, document the case. Emotional impact is human terrain: hear what actually drives the person. That is the fuel of persuasion, and it cannot be generated. Only recognised.

🤖 Agents do the outreach, humans do the close

Here is the split that works in 2026. The early steps, problem and pain, are research and first outreach. An AI agent scales that, for a thousand contacts at once, each one individual and personal.

But that only works with context. An agent without your company's business context writes generic smoke. Garbage in, garbage out. That is why you need the Context Engine, your company's business context as the foundation for the agents. With it the agent hits the real pain, not a platitude.

The late steps, pivot and close, stay human. Trust forms between people. The agent opens the door for a thousand contacts, the human walks through the few that matter. That is how a Hyperlean team plays big: a few pros plus AI agents, three instead of thirty.

A DACH footnote on scaling: cold email in Germany and Austria generally requires consent, and Swiss unfair competition law sets limits too. So the agent researches signals and prepares context. The outreach runs on channels that are allowed: LinkedIn, warm contacts, events, opt-in. What that looks like in practice is in the article on AI outbound on LinkedIn. And to make the parts play together, all of it belongs in a system for marketing and sales automation, not in isolated tools.

🎢 Outro: the psychology stays, the execution scales

✅ What shines. The wheel is timeless. Problem first, product last. That order holds with or without agents.

❌ What doesn't shine. An agent that starts at step six, straight at the product, sounds like every spam email. The order is not optional.

⚠️ Warning. Do not scale an empty pitch. An agent just makes a bad opening faster and wrong a thousand times over. The context has to be there first.

I said at the start: people buy the outcome, not the product. That does not change when an agent writes the first email. The six steps stay. What changes is only the pace and the scale. The psychology stays, the execution scales. The work sits right in between.