Conversational hypnosis is guiding someone into a focused, receptive, trance-like state using ordinary language, in an ordinary conversation, with their eyes open (and maybe with their coffee still in their hand). No swinging watch. No “sleep.” Just talk that’s built a certain way.
It’s a real craft, and it has real limits. I’ll show you both, including where the popular version oversells itself.
This article is about the mechanism: how the language actually works, and a careful accounting of what the research does and doesn’t support.
What conversational hypnosis actually is
Conversational hypnosis is the practice of eliciting trance through language, inside a normal-seeming conversation, without a formal induction and usually without the other person labeling what happened as “hypnosis.”
Trance here isn’t unconsciousness. It’s absorbed attention. You’ve felt it. You miss your highway exit because you were deep in thought. You get lost in a film and forget you’re in a seat with sticky floors. Attention narrows, the internal critic goes quiet, and suggestions land more easily than they would in ordinary chatter.
That absorbed state is the doorway. Conversational hypnosis is the craft of opening it with words instead of a watch.
Here’s the analogy I keep coming back to. Formal hypnosis is like taking someone into a soundproof recording booth: quiet room, clear ritual, everybody knows a session is happening. Conversational hypnosis is like getting a good take while the band is jamming in the garage. Messier. Eyes open. But the recording still happens, and often the person is more relaxed because nobody announced “we are now doing The Serious Thing.”
Now the myth. “Covert” makes people picture a puppeteer pulling strings. Real conversational work is far less dramatic and far more collaborative. You’re not overriding anyone’s will. You’re lowering the friction on ideas the person is already open to, in service of what they came to you for (if you’re a therapist). A subject who fundamentally objects to a suggestion tends to reject it, in trance or out. Some part of them stays on watch. That’s a feature, not a bug, and any good practitioner is glad it’s there.
So the working definition: trance-through-talk, eyes open, in service of the client’s own goal. Hold that. Everything else in this article is detail hanging off it.
The name soup: conversational, covert, indirect, Ericksonian
If you’ve shopped this topic, you’ve drowned in near-identical labels. Let me untangle them, because the confusion is doing real damage to how people learn this.
Conversational, covert, and indirect hypnosis are basically the same thing. Three marketing names for one idea: using artful, permissive language to elicit trance and deliver suggestion inside a conversation. “Conversational” describes the setting. “Covert” describes the fact that it isn’t announced. “Indirect” describes the language style, suggesting sideways rather than commanding head-on. Same animal, three angles. On this site I treat them as synonyms.
Ericksonian hypnosis is a different size of thing. Bigger.
Ericksonian hypnosis is a whole clinical school built on the work of Milton H. Erickson: formal trance inductions, therapeutic metaphor, utilization, confusion technique, task assignments given as homework, naturalistic therapy tailored to one specific human being. Conversational hypnosis is ONE BRANCH growing off that trunk.
When I say conversational hypnosis, I mean the language approach: conversational, covert, and indirect, with the Milton Model as its engine. Ericksonian hypnosis is the wider clinical school it grew out of. If you want that broader picture, Ericksonian hypnosis is where to go next.
Where did it come from? Meet Milton Erickson
Every road in this material leads back to one man, so let’s meet him without the halo.
Milton H. Erickson was a working psychiatrist in the mid-twentieth century, and he was genuinely gifted at reaching people other clinicians had given up on. His signature idea was utilization: whatever the person brings, you use it. Resistance, symptoms, an odd belief, a nervous habit. Instead of fighting it, you fold it into the work.
Utilization is the whole game, so let me make it concrete.
The classic story: a hospital patient insists he’s Jesus. Rather than argue him out of it, Erickson notes that Jesus was a carpenter, and puts the man to work in the hospital’s carpentry shop. He didn’t debate the delusion. He used it. Change the analogy for your own life: it’s like a sailor who can’t fight the wind, so he sets the sail to use it. The wind was going to blow either way. Utilization decides which way you point the boat.
Erickson also individualized relentlessly. He didn’t have one induction he ran on everybody. He tailored language, pacing, and metaphor to the specific person in the chair, which is exactly why “just memorize these scripts” misses the point of what he actually did.
Now, the mythology. Erickson’s life got wrapped in a heroic origin story: polio as a teenager, teaching himself to notice tiny movements while paralyzed, dyslexia, colorblindness, a mind rebuilt through sheer observation. Some of that is true, some is embellished, and a lot of it has hardened into folklore that’s tough to verify. I’d gently separate the method from the mythology. You don’t need the legend to be word-for-word accurate for utilization to be a brilliant, usable idea. Judge the technique on the technique.
One more thing worth flagging. Erickson’s clinical results come to us mostly as case reports and teaching stories, not controlled trials. That doesn’t make them worthless. Case reports are where good clinical ideas usually start. It does mean we shouldn’t wave them around as proof, and we should keep them separate from the controlled hypnosis research we’ll get to shortly. His primary writings, like Hypnotic Realities with Ernest Rossi in 1976, are where the language model was first laid out in detail.
For the fuller Ericksonian school, his life, the therapeutic approach, the broader techniques, see Ericksonian hypnosis. Here we follow the branch he’s most famous for: the language.
The language mechanism: the Milton Model
How does talk alone move someone into trance? One answer is the Milton Model.
The Milton Model is a catalog of language patterns that Richard Bandler and John Grinder pulled out of Erickson’s speech and organized into teachable rules. Where a lot of communication training tells you to be precise, the Milton Model does the opposite on purpose. It’s artfully vague. Loose, permissive, comfortably ambiguous language that gives the listener’s own mind room to fill in the meaning.
Why would vagueness help? Because a specific instruction gives the critical mind something to argue with. “You are getting sleepy” invites “no I’m not.” But “and you might begin to notice a certain comfort settling in, in your own way, in your own time” is almost impossible to disagree with. It’s so open that any experience the person has fits it. The listener does the work of making it true, and self-generated experience is stickier than anything you impose.
That’s the engine. A few pattern families make it run:
- Permissive, vague language. “You may find,” “perhaps,” “in your own way.” Soft frames that are hard to reject because they claim so little and permit so much.
- Presuppositions. Ideas smuggled in as assumptions rather than claims. “Before you drift deeper” presupposes that drifting deeper is happening. You didn’t assert it. You built it into the sentence.
- Embedded commands. A direct suggestion tucked inside a larger, innocent sentence and often marked out by a small shift in voice, so it reaches the listener without ever standing up and announcing itself as an order.
- Pacing and leading. First you describe what’s demonstrably true (you’re sitting, you can hear my voice, you’re breathing). Each true statement earns a small yes. Then you lead toward the suggestion, riding the momentum of agreement.
I’ve only sketched the families here. If you want the actual patterns with worked examples, go deep on nominalizations, the words that quietly move people, on how presuppositions do their work, and on multi-layered communication, which is where embedded commands live. Those articles teach what this section only names.
The core moves, framed
Below are the workhorse techniques of conversational hypnosis. I’m framing each in a paragraph so you see how the pieces fit. Each one has its own full article behind it, because each deserves more than a paragraph once you’re ready to practice it.
Pacing and leading
Pacing and leading is the backbone move, and everything else leans on it. You “pace” by matching the person’s current reality: their posture, their pace of speech, and above all statements they can’t help but agree with. Agreement builds a kind of rapport-momentum, like walking in step with someone until the stride becomes automatic. Then you “lead,” nudging toward where you’d like to go. Lead too early, before you’ve paced enough, and it doesn’t work. Get the ratio right and the transition feels like the person’s own idea.
Utilization
Utilization is Erickson’s crown jewel, carried into conversation. Whatever shows up, you use it. A car alarm goes off mid-session? “And as that sound fades, you can let yourself drift further.” The client fidgets, doubts, cracks a joke? All of it becomes material instead of an interruption. It’s the sailor setting his sail to the wind that’s actually blowing, not the wind he wishes he had. Utilization is what separates someone reciting a script from someone actually doing hypnosis, because reality never follows the script and utilization is how you stay with the person anyway.
Embedded commands
An embedded command is a suggestion hidden in plain sight. You take a direct instruction, “relax now,” and nest it inside a longer, casual sentence: “you don’t have to relax now before you’re ready.” A subtle shift in tone or a tiny pause marks the command out, so a part of the listener registers it while the conscious mind stays with the surface sentence. Done heavily it sounds ridiculous and everyone notices. Done lightly it’s one of the most useful tools you own. This one lives inside multi-layered communication, and it rewards restraint.
Presuppositions
Presuppositions move the argument before it starts. Instead of asking whether something will happen, you build it into the sentence as a given and ask about something smaller. “Would you like to go into trance sitting up or lying down?” quietly assumes trance is on the way. The only open question is the furniture. It’s the same structure a good server uses with “still or sparkling?”, never “do you want water?” You can spend a career refining these. Start with our breakdown of NLP presuppositions, then watch how often you already use them without noticing.
Double binds
A double bind offers a choice where every option leads the same direction. “You can go into trance quickly, or slowly, whichever suits you.” Both roads arrive at trance. The illusion of choice reduces resistance, because the person is busy deciding how rather than whether. Ethically this one deserves a careful hand, and we’ll get to that, because the ethical version binds toward a goal the client already signed up for, never toward one you’ve decided for them.
Two more language flavors worth knowing sit slightly to the side of the core five. Artful ambiguity, where a phrase can be heard two ways at once, so a second meaning slips past the guard. You can see it dissected in our pieces on mondegreens and oronyms and on how ambiguity carries indirect suggestion. And reframing language, the conversational judo of changing what an event means, which our Sleight of Mouth redefine pattern walks through. When resistance shows up and you want to move around it rather than through it, our guide to bypassing resistance with hypnotic language is the next stop.
Does it actually work? What the research says
Now for what the evidence actually supports, and what it doesn’t. This is the part that could sharpen how you practice.
It’s common to think indirect suggestion works better for “resistant” clients, or for people who don’t hypnotize easily. Good evidence for that is thin. Lynn, Neufeld, and Maré reviewed the direct-versus-indirect question and found the well-controlled studies don’t show indirect outperforming direct, and later work testing it on low-susceptibility subjects found the wording made no difference; the person’s own susceptibility did. What seems to move the needle is someone’s suggestibility and expectancy, more than how you phrase things.
That isn’t a knock on indirect language. It works — in a controlled trial, indirect, metaphor-based suggestion produced a real, measurable response. It just isn’t the secret weapon it’s sometimes made out to be. You reach for it because it lowers resistance and fits certain people and certain moments, not because it hits harder than plain speaking.
And from a practitioner’s standpoint, here’s what I do and how I think.
It ‘s likely that some people respond better to indirect suggestion and some respond better to direct suggestion. You can often judge by how they react to your suggestions in the really pats of a session. “Have a seat.” Do they bristle, or look uncomfortable? Maybe indirect is the way to go. It’s also quite likely that people respond differently in different phases of a session.
Based on my idea of how hypnosis and inductions work, I tend to favor indirect near the beginning of a session, and direct after an induction. And the deeper they go, the more direct I get. Why?
Indirect language is less obvious and less likely to have that mental guard at the gate stand up and say, “Hey, that’s not what I think.” You’re just telling a story about your grandfather or an elf.
After an induction, you’ve bypassed that critical guard at the gate. No need to be obtuse. Of course, it’s best to watch the client’s responses and judge for yourself, and I usually use a mix of both in a session. I just lean a bit more toward direct after an induction.
Ethics and consent for practitioners
A topic with “covert” in the name has to answer the ethics question head-on, so let’s do it without the hand-wringing and without the mind-control cringe.
The line I use is simple. Influence serves the other person’s goal. Manipulation serves yours. Same techniques, opposite direction. A double bind that helps a client who asked for help to stop smoking pick a quit date is influence. The identical structure aimed at getting someone to buy something against their interest is manipulation. The words don’t change. The intent and the beneficiary do.
So the first ethical question is “whose goal am I serving?”
This is why I keep hammering the client’s stated goal. Conversational hypnosis is influence toward what the person already told you they want. It’s not a license to install outcomes they never asked for. If a client came to you about performance anxiety, your covert language works on the performance anxiety. It doesn’t wander off to reshape their politics or their relationship because you happen to have opinions. The moment you’re steering toward an outcome the client didn’t request, you’ve crossed from practitioner to something worse, regardless of how skilled the language is.
Two more guardrails I’d hand any student.
First, remember what the research already told us: a subject who genuinely objects to a suggestion tends to reject it. That’s your built-in safety rail. Conversational hypnosis lowers friction on ideas the person is open to. It doesn’t bulldoze real objections.
Second, “covert” refers to the technique, not to a lie about the relationship. In a professional setting people have consented to be helped by a hypnotist. You’re not misrepresenting who you are or what you’re for. There’s a wide gap between not narrating your every linguistic move and actively deceiving someone about the nature of the interaction. Stay on the right side of it and this work is as ethical as any other form of skilled help. You don’t have to earn trust by announcing “presupposition incoming.” You earn it by only ever pointing this at what the client came to you to change.
How to actually practice this
Here’s how I’d train if I were starting over.
Drill one pattern at a time. Take presuppositions. Only presuppositions. For a week, in ordinary conversation, plant them and watch what happens. The next week, embedded commands. Nothing else. You’re building single patterns into reflexes, the way a musician drills one scale until the fingers own it and the conscious mind can let go.
Fluency in one pattern beats clumsiness in five.
Calibrate to the actual person in front of you. Watch the response. Did their breathing slow? Did the eyes defocus? Did they go quiet and thoughtful, or did they tense up? The person is telling you constantly whether you’re pacing well or leading too soon. Conversational hypnosis is a feedback loop, not a monologue. If a lead doesn’t land, you paced too little. Drop back, pace more, try again.
Let me put you in the chair. Client Example #1: a coaching client, call her Dana, keeps saying she “can’t relax.” Novice move: tell her to relax. She’s already told you that fails. Utilization move: “That’s fine, you don’t have to relax at all. You can stay exactly as alert as you are right now, and just notice, with all that alertness, what your hands are resting on.” You paced her reality (staying alert), used her resistance instead of fighting it, and slipped in a small lead toward internal focus. Nine times out of ten Dana’s shoulders drop within a minute, and she’ll swear she did it herself. She did. You just set the sail.
Now the common failure modes, so you can catch yourself.
- Leading too early. Impatience is the number one killer. You pace twice, get excited, and lunge for the suggestion. Pace more than feels necessary, then lead.
- Over-egging the embedded commands. Big theatrical tone shifts make people laugh or bristle. The mark of a pro is how little they telegraph. Subtlety, not drama.
- Sounding like a hypnosis robot. If your voice drops into that “hypno-voice” too suddenly, you’ve announced yourself. Stay yourself. Stay conversational. Ease into it.
- Ignoring feedback. Running your plan while the person’s face says “no.” Watch them, not your script.
Practice on low-stakes conversations first. The coffee order. Small talk. You’re not manipulating anyone by getting fluent with permissive language over a sandwich, and it’s a safe place to build reflexes before you’re sitting with a real client who needs real help.
If you want to go deeper on the actual patterns, the exact language, worked line by line with examples you can steal and adapt, that’s what our Hypnotic Language Mastery Pack is built for. It’s the patterns in this article, expanded into drills. No pressure. The articles linked throughout this page will take you a long way for free, and you can always come back for the full toolkit when you’re ready to get serious about the language.
Frequently asked questions
Is conversational hypnosis real?
Yes, with a caveat about what “real” means. Trance itself, the absorbed, focused state, is real and well documented, and hypnosis broadly has strong evidence: a 2024 review found 99.2% of hypnosis outcomes were positive. What’s oversold is the covert-superpower framing. Conversational hypnosis is a genuine skill for eliciting that state through talk. It is not mind control, and it won’t likely force someone past a genuine objection.
Is conversational hypnosis manipulation?
It can be, and so can any communication skill. The ethical line is direction: influence serves the other person’s stated goal. It’s good manipulation. Bad manipulation serves your goals at their expense. The techniques are identical either way. Used ethically, conversational hypnosis lowers resistance to ideas the client already asked for help with. A subject who genuinely objects to a suggestion tends to reject it, which is your built-in safeguard against the mind-control fantasy.
Can anyone learn conversational hypnosis?
Pretty much, yes, the same way most people can learn an instrument. The patterns are teachable and the mechanism isn’t mystical. What separates good practitioners from beginners is calibration, reading the person’s live feedback and adjusting, which only comes from practice. Drill one pattern at a time until it’s a reflex, then add the next. Reading about it builds no skill on its own. Reps do.
Is conversational hypnosis the same as NLP?
Overlapping, not identical. NLP formalized Erickson’s language into the Milton Model, which is the engine of conversational hypnosis, so the language side is shared ground. But NLP is a broader grab-bag of techniques.
Does indirect suggestion beat direct suggestion?
No, and this is the field’s most persistent myth. The best-controlled research, reviewed by Lynn and colleagues in 1993, found indirect Ericksonian suggestions are not superior to direct ones. In the pain research, what moved results was direct suggestion plus the subject’s suggestibility. Use indirection because it fits certain people and lowers resistance, not because it’s inherently more powerful. Often a clear direct suggestion to a receptive person does just as well.
References
- Thompson, T., et al. (2019). The effectiveness of hypnosis for pain relief: A systematic review and meta-analysis of 85 controlled experimental trials. Neuroscience & Biobehavioral Reviews. https://pubmed.ncbi.nlm.nih.gov/30790634/
- Lynn, S. J., Neufeld, V., & Maré, C. (1993). Direct versus indirect suggestions: A conceptual and methodological review. International Journal of Clinical and Experimental Hypnosis. https://pubmed.ncbi.nlm.nih.gov/8468104/
- Montgomery, G. H., et al. (2002). The effectiveness of adjunctive hypnosis with surgical patients: A meta-analysis. Anesthesia & Analgesia. https://pubmed.ncbi.nlm.nih.gov/12032044/
- Rosendahl, J., et al. (2024). Efficacy of hypnosis: A 20-year meta-analytic review. Frontiers in Psychology. https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2023.1330238/full
- Sturt, J., et al. (2012). Neurolinguistic programming: A systematic review of the effects on health outcomes. British Journal of General Practice. https://pmc.ncbi.nlm.nih.gov/articles/PMC3481516/
- Erickson, M. H., & Rossi, E. L. (1976). Hypnotic Realities. Irvington Publishers. (Primary source for the Milton Model and utilization.)
Enjoy,
Keith.
