The Hour After Your Last Client

Sooner or later it comes up in the room.

A client mentions, half apologetically, that they have been talking to one of those apps at two in the morning, and that it helped. You keep your face where you keep it, and you notice three things at once. That it probably did help. That there is a question underneath the sentence, which is whether you are still necessary. And that the app has no idea what next Thursday is the anniversary of.

That is one of the two loud stories, arriving in the gentlest form it takes. The rougher form comes as a forward from a colleague, under a headline ending in a question mark, which is generally how you can tell that nobody knows.

The other story wears a lanyard. There is a booth at the state conference and an email every few weeks afterward. Notes written before you stand up. Intake handled. Caseload full by spring. The clinicians who move first will be the ones still standing, and the first thirty days are free.

Both belong to a year nobody has practiced through yet, and neither touches anything on your desk tonight. Four notes are still open from Thursday. A denial from six weeks ago is under the mail. The inquiry that came in on Monday has gone three days without an answer.

Neither of those is this book’s business. The four notes are.

Take the story where the profession ends first. It settles fast, and everything that settles it happens in a room with the door shut.

It cannot hold a silence. Twenty seconds after somebody says the thing they have never said out loud to anyone, and the whole treatment is whether you speak or you don’t. That is a decision made in real time, by a person, with another person watching their face. A language tool has nothing to do in a silence.

It cannot hear what is not being said. Three weeks running, the marriage gets walked around. The joke that turns up every time the father does. The breathing that changes a beat before the crying starts. None of that is in a transcript, and a transcript is the most a machine will ever have.

It cannot be the relationship, and the relationship is the treatment. You were taught that the alliance does more of the work than the model does, then spent years watching that turn out to be true. It cannot be disappointed in and then forgiven. It cannot be the first person somebody has ever been angry at without losing.

It cannot sit with risk. When somebody tells you, forty minutes in, what they have been thinking about at night, the next ten minutes belong to a licensed human being with a phone, a plan, and a duty.

It cannot be trusted with a word of any of it. That is not a limitation, it is a wall, and the wall goes up before this book offers you anything. Nothing a client has told you goes into a consumer chat window. A paragraph on that shortly, and a chapter later.

It carries no license. Yours has a number on it and a board behind it that can take it away. Your notes are legal records that can be subpoenaed and read aloud in a room full of strangers.

And it does not know your rules. Ask what your code says about a review that appeared without your asking, and out comes an orderly paragraph about some other license in some other state. Chapter 10 handles that. Until then, treat anything it says about ethics or boards as a rumor with good posture.

That is the answer to the first story.

The unease underneath is not foolish. Apps that talk exist. They are free, they are awake at two in the morning, and they will get better at talking, because producing sentences was always the cheap part.

What that changes is not the hour. It is the ten seconds before the hour: where a frightened person looks first, and who they find. A page that answers the question they actually typed. Whether the practice they emailed at eleven at night replies on Tuesday or on Friday. Which argues for reading on: the clinicians hardest to route around are the findable, specific, fast ones, and Chapters 7 and 9 are about nothing else.

Now the part a softer book would leave out. A book that told you nothing was moving would be caught at Chapter 4, and then you would stop trusting the useful pages.

Something is moving. None of it is clinical.

Graduate school covered theory, ethics, assessment, and supervision. Nobody sat you down and mentioned that a private practice is a small documentation business with a therapist attached.

So: the notes. Treatment plans due for review, written at ten at night in language you have retyped several hundred times. The intake packet you assembled in 2019 out of three borrowed PDFs. The prior authorization. The superbill explanation you deliver out loud, weekly, to people whose eyes glaze at the word deductible. The inquiry reply you write three times, because too eager reads as desperate and too formal reads as cold. The fee increase you have deferred three years running, because the letter feels impossible.

None of that is therapy. None of it gets billed. Nearly all of it happens after the last client leaves, or on a Sunday afternoon, or not at all.

That pile is the part these tools genuinely handle. Not the clinical judgment inside a treatment plan; the eight sentence stems around it that you rebuild every time. Not the decision to raise your fee; the ninety words that let it land.

Here is the fence, in full, because in your profession it comes before the benefits rather than after them. Nothing a client has told you goes into one of these tools. Not the session. Not the note you were going to tidy up. Not the story with the name removed, because the name is the least identifying thing about a person and a recognizable story is a breach without it. Not on a paid tier, and not just this once at ten at night when you are seven notes behind and it would take four minutes. The company behind a consumer account has signed nothing about the people you treat. A separate category of clinical documentation tools works under signed agreements, and Chapters 10 and 11 cover them. Every workflow in this book was built to run without a single client in it, which leaves nearly all of them standing.

These tools also invent, in precisely the tone they use when they are right. They will hand you a CPT code that is nearly correct, an appeal deadline that appears in no plan document, and the whole wellness-brand vocabulary of promised outcomes your advertising rules do not let you publish. So the workflows here split the same way. The facts are yours: codes from your billing records, deadlines from the denial letter, dollars from your books. The sentences are its half. Where that is easy to fumble, a WATCH OUT box gives you the check, and it takes about a minute. Numbers, promises, rules.

One commercial sentence, said once. A person in trouble emails three therapists in a single sitting, usually from a phone, usually late. The one who answers the same day, warmly, with the fee stated plainly and one clear next step, generally gets the consult call. That was always true. What changed is the price of the reply: it used to cost an evening you did not have, and now it costs about four minutes between a two o’clock and a three o’clock.

An ordinary Thursday, then. Nothing dramatic in it, which is the point.

Six clients. The ten-minute gaps went on the bathroom, the coffee, and one of them on somebody who needed to run over. The last hour ends at ten past six. You mean to write the notes now, while they are fresh.

The first two are good. Specific, defensible, ten minutes each. The third is thinner. By the fourth you are typing a sentence you have typed six hundred times and are aware of typing it, and at twenty past seven you stop, because you have not eaten since noon. Four notes stay open.

The denial has been on the desk for six weeks. You have picked it up twice and put it down twice, not because you agree with it, but because an appeal letter is a genre, formal and citation-shaped, and nobody ever taught you the genre. There are four more like it in the drawer.

Then the inquiry from Monday. She described a panic attack in the first paragraph and apologized twice for the length. You read it standing up between sessions, and composed the reply in your head walking to the car: warm, clear, honest about the fee. On Friday she books somebody else, and you never find out that it was not the fee.

Nothing in that evening was a competence problem. The two notes you wrote at six were excellent, and the reply you composed in the car was the right reply. The trouble was that it was in the car. The arithmetic is the whole story: a good note costs ten minutes, a good inquiry reply costs twenty, and Thursday handed you six clients and one of you. The four open notes then join the stack that owns Sunday.

So here is the book, stated plainly, so you can decide now whether to keep going.

This is a manual for aiming one everyday AI chat tool, any of ChatGPT, Claude or Gemini, at the business side of a practice. Note templates and treatment-plan language banks, built once and used for years, with every word of clinical content staying in your own records. Intake packets, policies, and consent outlines that reach your attorney as drafts rather than as an unstarted task. Inquiry replies, the fee conversation, the between-session boundary reply. Superbill explainers, appeal letters, balance notes that leave the alliance intact, the rate increase you keep putting off. Referral letters, directory profiles, handouts, worksheets, group curricula. Marketing that fills a caseload without going near a testimonial. There are 150 prompts, numbered, and a library at the back that takes about ten seconds to search.

Each of the 150 was tested in all three of those tools, and any prompt that needed one particular tool to behave did not make the book. So this is not a forecast about AI and therapy. It is the wording that survived.

It is not a note writer, a clinical tool, an ethics ruling, or a substitute for your board, your association, your liability carrier, or an attorney. It will not tell you how to do therapy, and nothing in it asks it to. And it never asks you to put a client near a chat window, which is the sentence to keep if you keep only one. Chapter 10 gathers every RULES CORNER box into one place; read it in your first week.

None of this asks you to be good with computers. A phone, an email address, and half an hour, which Chapter 2 spends for you. Chapter 3 holds the one skill that matters, and you have used it since your first month in practice, because briefing one of these tools is making a referral. Who the person is. What is going on. What would fit. What to watch for.

One word about size before you start.

Nothing here changes your life, and nothing here changes your practice. It changes the hour after your last client, and inside a month, most of your evenings and a fair share of your Sunday. The page opposite counts it chapter by chapter and comes to about eleven hours in a normal week. It is a real number and it is not a gift. Your first week runs slower than doing the work by hand, and around week three it starts paying you back.

Two things out of twelve chapters is the usual haul, and it is the correct one. Most clinicians take the inquiry kit and the Friday appeal habit, never open the marketing chapter, and get four hours back. There is no prize for finishing the book.

Nobody is coming for the chair. The twenty seconds of silence after somebody finally says it, the noticing of what got walked around again, the person who gets to be furious with you in week nine and finds you still there in week ten: that is the treatment, and none of it is a writing task. What is on offer is more modest than either story you were handed, and much more use. The note written in the ten minutes it belongs in. The inquiry answered the same evening. A Sunday that is not spent paying for Thursday.

Half an hour, whenever you next have one.