Copy, Paste, Done: My AI Report Method
From Six Pages to Six Minutes: How AI Changed My Reports
If you read The Report Writing Rabbit Hole, you know I used to write six-page reports, and then I moved to a simple report sheet I filled out live at the patient's initial assessment. That sheet served me well — it was efficient, it worked for my practice needs at the time, and I still believe there's always room to think outside the box in how we run our practices. That includes how we document and share patient information — whether that's a “report” or any other practice system.
But I've since found something even more efficient. Since that post, I've started using AI to generate a formal report — created once, using what I've already documented, to be shared with the patient or their providers. Here's exactly how I do it, and what you need to know before you try it yourself.
How This Actually Works
This isn't AI listening in on your appointment and guessing at what mattered. It's simpler than that — and you stay in control of every word.
You copy and paste the data you already have — a health history form, an assessment note, a referral from a provider, whatever pieces you want reflected — into an AI platform like Claude. You tell it how you want that information organized. It fabricates the report in the structure you've designed.
Personally, my report includes:
A quick overview of health history as it relates to OMDs and any related concerns
A list of my observations from the assessment
Data points — measurements, tests passed or failed
My recommendation(s) for therapy and/or other interventions
That's it. That's the whole report. No mystery, no six-page marathon.
Stop Recording the Same Data Three Times
Here's the part that used to drive me a little crazy, and I bet it's driving you crazy, too.
A patient fills out a health history form. You read it, and you may jot some of that information down again at the first visit. Maybe you even have some of it noted from a previous phone call. Then you write it into your appointment notes. Then you write it again into the report. That's the same information, traveling through your hands several times. It's not more thorough — it's just more work.
Reframe it: the patient fills out their health history once. That data lives in their file as a clinical record. When it's time to write the report, you're not re-entering it — you're referencing it. And if you're using AI for that step, referencing it is a copy and paste.
You did the work of collecting the data already. Let the report-writing step be exactly that light.
Who Is This Report Actually For?
This is the same question I asked myself in the last post, and it's worth asking again here — because it's what shapes your AI template.
Do you want the report to read clinically, or in plain, layman's terms? Brief enough for a quick once-over, or a more involved narrative? Written for the patient, a referring provider, or both?
The beauty of building a template that produces what you want is that you only do the setup work once. You test it, see what it gives you, tweak it if something's off — or if your preferences change down the road — and then you use it again and again. You're not rebuilding this every time you sit down to write. You're running the same trusted process on new data.
Not only does this make your system flow seamlessly — it's organized, it looks professional, and you always know exactly where to look. Providers know what to expect and get the impression that you know what you're doing.
The HIPAA Conversation
This is the part I don't want you to skip.
One of the biggest concerns I hear is that non-healthcare-focused AI platforms aren't encrypting your data — they're collecting it. Unless you're on an expensive HIPAA-compliant tier, that data isn't protected the way your patient information needs to be.
Which means: don't put any identifying patient data into these platforms at all. Omit the name and date of birth entirely. Leave a blank space in your template where that information belongs, and fill it in manually afterward, inside your own HIPAA-compliant system.
Your clinical content — the health history summary, the observations, the recommendations — isn't identifying on its own. The name and birthdate are what tie it to a real person. Leave those out of the AI conversation entirely, and you've closed the gap.
The Other AI Way: Scribes
Copy-paste isn't the only path into AI-assisted reports. AI scribes — tools that listen to your appointments and generate notes for you — are worth knowing about, even if you don't use one.
Most people think of scribes as a SOAP-note tool, but some let you customize the output format entirely, including into a report structure. Heidi and BastionGPT are two worth exploring — both support custom templates beyond the standard note format, and some can also accept uploaded documents, not just live audio, which puts it a bit closer to the copy-paste method above and collecting all the data you want included in your reports.
Check your EHR too — many have a scribe add-on built in. As of this post, Jane App supports templates outside of standard SOAP notes; SimplePractice does not.
One real limitation to know before you rely on a scribe: its job is to listen and record. If a detail matters to your report but was never said out loud in the session — something from an old health history form, a prior note, a referral letter — a scribe won't know it exists unless you say it out loud during the session, or the platform lets you feed it in separately. That's the trade-off. The copy-paste method puts every source in front of the AI on purpose; a scribe only knows what it heard.
The Whole Process, In Four Steps
Decide on a format
Build your template
Copy and paste your data in (no names, no dates of birth)
Download the report, fill in the identifying details for the official record (only to be stored in an encrypted space, of course)
That's it. That's the whole system.
Come Build One With Me
Want to see it in action? I walk through this whole process step by step in the Pop-Up Workshop, and the recording is available on demand. Get it here.
