CarePilot works best when you think of it less like software and more like a sharp third-year resident sitting in on your visit — someone with real clinical knowledge, but zero context beyond what you actually tell them. This article covers the first of two mental models we use constantly at CarePilot: it can hear you, but it can't see you.
The mental model: CarePilot has the clinical knowledge of a sharp third-year resident, but none of the context. Everything it knows about this visit comes from what it hears — nothing else.
CarePilot Is Listening on Speakerphone
Picture a sharp resident sitting in on your visit over a speakerphone from the next room. They can hear every word you and the patient say. They cannot see the patient's shoulder. They cannot watch you perform Neer's test. They cannot read the vitals monitor over your shoulder. If you don't say it out loud, as far as that resident's chart is concerned, it never happened.
CarePilot works exactly the same way. It can hear you. It cannot see you.
This matters most during the live visit, on the Note tab — the note, the diagnoses, and the billing codes are only as good as what got verbalized in the room.
What to Verbalize
Objective exam findings — anything you see, feel, or read that isn't spoken to the patient directly
Special tests and their results — name the test and the result out loud, even when it feels redundant
Vitals or values you're reading off a screen — CarePilot can't see the monitor
Your clinical reasoning — why you're ruling a diagnosis in or out, not just the diagnosis itself
Plan rationale — why this medication, this dose, this referral, not just the order itself
Weak vs. Strong Verbalization
Left unsaid (CarePilot never hears this) | Said out loud (CarePilot captures this) |
Silently palpating the shoulder and moving on | "Tenderness to palpation over the greater tuberosity and subacromial region." |
Doing Neer's and Hawkins-Kennedy without narrating | "Neer's impingement sign is positive. Hawkins-Kennedy is also positive." |
Jumping straight to a diagnosis | "This fits rotator cuff tendinopathy rather than adhesive capsulitis, because passive range of motion is full and pain-free." |
Reading vitals silently off the monitor | "Heart rate 78, blood pressure 118 over 76 — both fine for the stimulant dose." |
The richer your narration, the more accurate CarePilot's diagnoses, coding, and orders will be. A vague or silent exam produces a vague or generic note — not because CarePilot guessed wrong, but because it genuinely never received the information.
It Applies to Everything You Want Captured
Physical exam findings, positive and negative
Differential reasoning — why you're ruling other diagnoses out, not just what you're ruling in
Patient education you hand over silently (a printed form, a handout) — say what you're giving them
Anything you'd normally chart from memory after the visit — say it in the room instead
Getting Started
No setup needed — this is a habit, not a feature. On your next visit, pick one thing you'd normally do silently (a special test, a vitals check, a piece of reasoning) and say it out loud. Watch for it in the generated note afterward.
Questions? Reach out any time in the Help Center chat.