The Language Pattern That Speaks Beneath the Conversation: Embedded Commands for Leaders

In 1997 I flew to Duluth, Minnesota, to sit in one of Richard Bandler's trainings at the Black Bear Hotel and Casino. I was a university student taking extra days off school, half-convinced I was wasting time and money on something I could not yet name. Two friends came with me. Devon, who has since passed on. Wiki Tay, who still trains and speaks today.

I came home and ran my first program with my then business partner Jerry. Seven participants showed up. Two of them were my parents.

At the end I asked each person what the biggest thing they had taken away was, expecting stories about anchoring or state control or the swish pattern. My mother said: "State management. Because you now can manage your emotions much better, and it is a way to become a better version of yourself."

My father sat with it for a moment. Then he said something I have never forgotten: "Finally, I have a way to put all the things that I thought of into a cogent framework, a language to communicate with people."

He did not mention one technique. He mentioned language. He had spent a life with a head full of ideas and no way to reach the people across the table from him, and what moved him was not a gimmick. It was the discovery that the words themselves could carry weight beneath their ordinary meaning.

That, more than any single tool, is what the Milton side of NLP gives you. Not louder words. Words that work under the level of the conscious mind, the way a current moves under a calm river.

This article is about one specific piece of that: embedded commands. What they are, how they work, where the line is between influence and manipulation, and how a founder uses them with an intact conscience.

An embedded command is a directive tucked inside a grammatically normal sentence, marked off by voice tone, a slight pause, or a gesture, so that the conscious mind hears the surface sentence while the unconscious mind registers the instruction underneath.

Here is the pattern. Say the sentence "I wonder why you keep delaying the decision on the new hire."

On the surface it is a question about your reasoning. But drop the voice on the words "keep delaying" and "decision" and it becomes two sentences stacked in one. The surface sentence acknowledges your dilemma. The embedded sentence quietly instructs. The instruction lands below the radar of the analytical mind, which is too busy parsing the question to object to the order.

The mechanism rests on a simple observation about attention. When someone challenges your conscious reasoning, your guard goes up. When someone speaks in a way that does not demand agreement or argument, your guard stays down, and the language registers in a different way.

Milton Erickson spent a career exploiting this. He found that direct suggestions often met resistance, because people resist being told what to do. Indirect suggestions, delivered inside a story or a question or a description, slipped past that resistance because the conscious mind had nothing to push against. He called this the use of indirect forms of suggestion to make hypnotic responses more acceptable and more easily accepted.

What makes this important for a founder is not hypnosis in the stage-show sense. Most people never go into anything resembling a trance. What happens is subtler: the other person absorbs the frame you have placed, even while they argue with the surface words you used.

Here is the uncomfortable part. This is a real, teachable mechanism. And like every tool of influence, it can be aimed at genuine cooperation or at compliance you would be ashamed of if it were named out loud. The difference is not the technique. The difference is the frame of the person running it.

This is also where the honesty gate matters. A good deal of the writing around embedded commands is dressed up as a kind of linguistic wizardry that bends people to your will. It does no such thing. Modeling claims about the unconscious are drawn from Erickson's clinical observations and Bandler and Grinder's modeling of his work; they are not the output of controlled trials. Treat the technique as what it is: a real pattern of communication with real effects, and a pattern whose scope is far narrower than the marketing around it.

A conversation about influence almost always collapses into a question of intent. The frame matters more than the phrasing. And the frame is built well before you open your mouth.

Consider two versions of the identical embedded command. Version one: "As you sit here thinking about the renewal, you can notice how ready you are to commit." Version two sells a client something they do not need, using the same shape of sentence to rush a signature they will regret.

The words can be identical. The relationship is not. In the first case you are using language to help someone discover a decision they already have access to. In the second you are using language to extract a decision they have not genuinely made.

Here is the test I use, and it is the same test that separates a calibration skill from a manipulation tool: after the conversation ends, does the other person feel more capable, or less? Does the language expand their options, or does it narrow them for your benefit?

Embedded commands are at their most honest when they surface a resource someone already possesses. You are not planting a foreign idea. You are making it easier for them to hear an idea that is already theirs. The sentence "you will notice how clear the decision feels once you stop rushing it" does not manufacture a decision. It gives permission for a decision that is already forming in them to rise.

This is where the deeper point of the recent conversations around NLP shows up. In 2025 the debate about whether neuro-linguistic programming works kept circling the same ground on Reddit and in practitioner forums, with defenders insisting it works and skeptics demanding proof. The real split was never technique. It was epistemology, a question of what counts as evidence. People who benefit from NLP treat internal experience as data. Skeptics demand external proof before they will treat internal experience as data. Both are running different operating systems, and neither conversation will ever satisfy the other until that is named.

The same divide runs through the ethics question. If you believe influence is only legitimate when it is overt and fully reasoned, embedded commands will always look like a trick. If you believe that human beings are influenced by language whether they know it or not, and that the only choice is whether that influence is run skilfully and honestly or clumsily, then the question shifts. It is no longer "may I influence." It is "what am I influencing toward, and does the other person come out ahead."

You are not going to stop influencing people with embedded commands. You do it already. Every sentence you place carries an implicit instruction to see the world a certain way. The real decision is whether you run that influence on purpose, with a frame you can defend, or leave it running blind.

Building this well takes supervised repetition, the same way every other precision skill does. Reading about it fills your head with concepts. Running it in real conversations, with feedback, builds the calibration. Here are the three moves that matter most.

The first move is to notice the natural voice markers you already produce. When you want to emphasize a word, you change your tone, you slow down, you add a beat. You do this all day without thinking. An embedded command asks you to do exactly that, on purpose, on the words you want to land underneath. The sentence "I wonder how soon you will begin" becomes two sentences when "begin" falls in a lower, slower register. The conscious mind hears the question. The intent underneath shifts.

The second move is to run the command inside a reason or a story, never as a blunt order. Erickson's genius was that he almost never issued direct commands. He told a story about someone else making a decision, and let the listener's unconscious attach itself to the path. As a founder you do not need to become a storyteller of trance. You need one habit: before you state what you want, state the reason it makes sense, and let the reason carry the instruction. "The team will move faster on this quarter's numbers once they know the target is reachable" carries the embedded command to move faster, wrapped in a statement of fact that nobody feels the need to resist.

The third move is the one nobody talks about: the ecology check, run on yourself. Before you place a command, ask what would change in the other person's world if they acted on it. Are you making them more able, on their own terms? Or are you engineering an outcome they will feel had a hidden cost? If the answer to the second question is anything but no, do not use the pattern on them. Use it on yourself first, to find the honest version of the request.

Here is what that looks like in practice. A founder is about to speak to a hesitant investor. Instead of pressing with direct arguments, she opens with: "As you look at the numbers, you will see why this round is already decided by the market." The embedded command is not pressure to sign. It is a reframe that lets the investor's own analysis arrive at the conclusion. The investor feels smarter, not pushed. The founder gets the round. Both walk away enlarged.