> ## Content Index > Fetch the complete content index at: https://www.mindovermoney.ai/llms.txt > Use this file to discover other available public pages before exploring further. # Steal My Prompt Vol. 51: The Parking Lot Recall - URL: https://www.mindovermoney.ai/prompt-library/ai-prompt-to-remember-what-your-doctor-said/ - Published: 2026-09-15T11:00:49.000Z - Updated: 2026-09-15T11:00:49.000Z - Description: You leave the visit certain about some of it and fuzzy on the rest. This prompt takes a three-minute dictation from the parking lot, tags every item as heard, unsure, or lost, then builds your instruction list and a portal message that asks the office to confirm the rest. - Author: Santosh Savel - Tags: #prompt-library You walk out of the visit with a medication change, two instructions, and the question you meant to ask. By the parking lot, one of them is already fuzzy. The provider said it once, quickly, while the next thing was being said, and the after-visit summary that lands in the portal three days later will not mention it either. By the time you sit down to write it out, what you have is a memory you half trust. The usual move is to write down whatever you remember and treat all of it as equally true. That is the mistake. Some of it you heard clearly, some of it you are reconstructing, and some of it you know was said and cannot bring back at all. Those three kinds of memory should not be handled the same way. Talk into your phone for three minutes before you leave the lot, everything you remember, hedges included. This prompt sorts it. Each item gets tagged HEARD, THINK, or LOST, and the tag decides where it goes. HEARD becomes your instruction list. THINK and LOST become a short portal message that asks the office to confirm each one in writing. It also names the one item that should not wait for a portal reply. You drive home with a record you can act on and a request for the rest. Two weeks ago,[ Vol 49 reconciled notes from five providers](https://www.mindovermoney.ai/prompt-library/ai-prompt-to-coordinate-care-between-doctors/). This one rescues the conversation before the note exists. I am building this before I need it. My wife's visits get more frequent from here to the November due date, and I want each one documented into a single running conversation I can hand to the next provider, rather than a stack of half remembered instructions. ## From User to Builder **Get AI Workflows Like This Every Tuesday.* Subscribe for Free Email sent! Check your inbox to complete your signup. Neural Gains Weekly. No Spam. Unsubscribe Anytime. ****Explore AI Out Loud** Two people at different points on the same AI journey, talking through what they are learning. The ideas, the failures, and the breakthroughs. Launches September 15\. New episodes every other Tuesday. [Subscribe on YouTube ](https://www.youtube.com/@exploreAIoutloud?ref=mindovermoney.ai) ### **What You Can Use This For** 1. Any visit where a medication was started, stopped, or changed and you are not certain you caught the dose. 2. A prenatal appointment in the final stretch, when the plan changes visit to visit and you are the partner expected to remember it. 3. A parent's specialist visit you joined by speakerphone and only half heard. 4. A pediatric visit where you were managing the child and listening at the same time. 5. A follow-up after a test result, when the explanation came fast and the next steps came faster. ### **How to Use It** 1. Open Claude on your phone. The prompt also works in ChatGPT, Microsoft Copilot, or Gemini on the free tier. 2. Turn on the reasoning option before you paste. In Claude that is "Extended thinking," in ChatGPT the reasoning model, in Copilot "Think Deeper," and in Gemini "Deep Think." A slower read is better at telling a hedge from a fact. 3. Copy the prompt, fill in the two bracketed fields, and send it. It replies with one word and waits. 4. Do not go home first. Tap the microphone and talk for three minutes. Start with whether anything changed today, then say the rest in any order. Say "I think" and "something about" when that is the truth. The hedges are the input. 5. Read the first line, which is the item that cannot wait. Then send the portal message before you drive, while the office is still open. **Pro tip:** at the next visit, paste last time's Do List above the new dictation and ask for only what changed. One prompt becomes a running record of the whole stretch of care, in five minutes a visit. ### **The Prompt** *You are reconstructing a medical visit from memory, and your standard has three rules. Nothing goes on the instruction list unless I clearly heard it. Anything I am unsure about goes back to the office in writing. You never fill a gap with a plausible guess, because a confident wrong instruction is worse than a blank.* *The visit was with \[WHO THE VISIT WAS WITH AND WHAT IT WAS FOR\]. The patient is \[WHO THE VISIT WAS FOR AND YOUR RELATIONSHIP, SUCH AS ME, OR MY WIFE AT 34 WEEKS, OR MY FATHER\].* *When you have read this, reply with the single word READY and one line reminding me to start my dictation with whether anything changed today, meaning a medication started, stopped, or given a new dose, a referral, a test ordered, or a date set. Then wait. Do not ask anything else, before or after.* *What I send next is everything I remember, dictated right after the visit. It will be out of order. Where I say "I think," "something about," or "they mentioned," that is a real hedge and you treat it as one. If I never say whether anything changed, treat any change you infer as unconfirmed and route it to the portal message.* *After my dictation, break it into single items, one instruction, fact, or question per item, and give each one of three tags. HEARD when I stated it as fact with a specific attached, such as a name, a number, a date, or a direction. THINK when I hedged, stated it without a specific, or said two things that do not fit together. LOST when I said something was discussed and could not recall the content, or asked something and did not get an answer. Never upgrade a tag to make the result look more complete.* *Then produce four things, in this order.* *1\. THE ONE TO CALL ABOUT. One line naming the single uncertain item that should not wait for a portal reply, and why. If nothing rises to that level, say so in one line.* *2\. THE DO LIST. HEARD items only, each as one instruction with a when, ordered by the purpose of the visit. Nothing tagged THINK or LOST appears here, even if it seems obvious.* *3\. THE PORTAL MESSAGE. Written from me, addressed to the office I named, under 150 words, in plain language. Ask them to confirm in writing every THINK and LOST item as a numbered list, each phrased so a nurse can answer it with a yes, a no, or a single fill-in. Put any medication item first. Do not restate the HEARD items. Close by asking for the reply through the portal so the answer is on record.* *4\. THE RECALL TABLE. Every item, three columns, WHAT I REMEMBER | TAG | WHY THAT TAG. Any row involving a medication, a dose, or an instruction to stop something goes first, whatever its tag. This is my record of what the visit sounded like an hour after it ended, so keep my wording.* ### **Transparency and Notes** - Built in Claude on the free tier. Runs in ChatGPT, Copilot, and Gemini as well. Dictation uses the microphone button in each mobile app. - Model-agnostic. The ready-and-wait step is written into the prompt, so it behaves the same everywhere. - Privacy. Do not dictate the patient's full name, date of birth, or record number. A first name and the specialty are enough. Check your tool's data settings before pasting health details, and know that some workplace or school accounts retain chat history. - Educational only, not medical advice. If something in the room felt urgent, call the office. A portal reply can take days, and this prompt is built to tell you which item cannot wait for one. - [Vol 49, The Care Team Coordinator](https://www.mindovermoney.ai/prompt-library/ai-prompt-to-coordinate-care-between-doctors/), is the companion to this one for the weeks when the notes already exist.