AI SOAP Note Generator for Doctors

Generate SOAP notes from clinical conversations

SOAP notes remain the most widely used documentation format in outpatient medicine, and for good reason — the four-section structure organizes clinical thinking in a way that other clinicians can follow quickly. The problem is not the format itself but the time it takes to write one from scratch, especially after a complex visit.

Dictum helps reduce documentation burden by generating a SOAP note draft from either a recorded patient encounter or a post-visit dictation. The clinician records the conversation (or dictates a summary), and Dictum produces a structured note within minutes. From there, you review each section, correct anything the AI got wrong, add clinical detail it could not infer, and finalize the note for your EHR.

This is not a transcription tool that dumps a wall of text and leaves you to sort it out. Dictum actively maps conversational content to the appropriate SOAP section, producing a note that looks like something you would write — not a raw transcript with headers slapped on top.

What a SOAP note includes

For clinicians already familiar with the format, this is a quick refresher. For those evaluating documentation tools, understanding what each section expects helps clarify what the AI is doing and where it might need correction.

• Subjective — what the patient reports. Chief complaint, history of present illness, review of systems, medication list, allergies, social and family history as discussed during the visit.
• Objective — what the clinician observes or measures. Vital signs, physical exam findings, lab results, imaging reviewed during the encounter.
• Assessment— the clinician’s impression. Working diagnoses, differential considerations, clinical reasoning connecting the subjective and objective data.
• Plan — what happens next. Medications prescribed or adjusted, orders placed, referrals made, follow-up timing, patient education and instructions.

Each section builds on the one before it. A well-structured SOAP note tells a story: the patient said this, you found that, you concluded the following, and here is what you are doing about it. Dictum preserves this logical flow in the drafts it generates.

How Dictum structures each section

Dictum does not simply split a transcript at arbitrary points. It uses clinical language models to understand which parts of the conversation correspond to each SOAP section, based on the nature of the content rather than its position in the recording.

Subjective: When a patient describes their symptoms, medication side effects, or functional limitations, Dictum routes that content to the Subjective section. It also captures relevant history items — surgical history, family history, social factors — when they come up during the conversation, even if they are mentioned out of the typical HPI sequence.

Objective:If you state exam findings aloud (“lungs are clear bilaterally, no wheezes or crackles”), Dictum places those in Objective. The same applies to vitals, point-of-care test results, or imaging findings you discuss with the patient. Dictum does not fabricate exam findings — if you did not mention them, they will not appear.

Assessment:Your clinical reasoning — statements about likely diagnoses, severity impressions, or how today’s presentation connects to prior visits — goes into Assessment. This section often needs the most clinician editing because diagnostic reasoning is the hardest thing for an AI to capture accurately from conversation alone.

Plan: Treatment decisions, prescription changes, referrals, and follow-up instructions are captured in Plan. Dictum organizes plan items by problem when multiple issues are addressed in a single visit.

Example SOAP note structure

Below is the general structure Dictum produces. This is not a real patient note — it shows the kind of content that maps to each section so you know what to expect from the output.

Your actual note will contain the specific clinical content from the encounter. Dictum fills each section with the relevant details from the conversation, and you review everything before it goes into the chart.

Specialty-adapted templates

A SOAP note for a psychiatry follow-up looks different from one for an orthopedic consultation. The sections are the same, but what belongs in each section — and how detailed it needs to be — varies by specialty.

Dictum includes templates adapted to more than a dozen clinical specialties. A psychiatry template, for example, emphasizes mental status exam elements in Objective and treatment response in Assessment. An orthopedics template structures the physical exam around joint-specific findings and includes procedure-specific plan elements.

If the built-in templates do not match your workflow, you can build your own using custom clinical note templates. Custom templates let you define the subsections, ordering, and default content expectations for each SOAP section. Once saved, Dictum applies your template to every new note you generate.

Clinician review and EHR workflow

Every SOAP note Dictum generates is a draft. It does not go into any EHR automatically, and it does not leave Dictum unless you explicitly export or copy it. The review step is where the clinician adds their expertise — verifying that the content is accurate, clinically complete, and appropriate for the medical record.

Clinicians should review AI-generated documentation before adding it to the medical record and should use Dictum in accordance with their organization’s policies and applicable laws.

The editing interface lets you modify text inline, reorder content between sections, and add information the AI missed. When the note meets your standards, you copy it to your EHR or use Dictum’s export options. The process is designed so the AI handles the initial structuring — the part that takes time but does not require clinical judgment — while you handle the parts that do.

Dictum works alongside your existing EHR, not as a replacement. Whether you use Epic, Cerner, Athena, or another system, the output is plain text that pastes into any note field. For clinicians using the ambient AI medical scribe workflow, the SOAP note is generated directly from the encounter recording. For those who prefer to dictate, the same structured output is produced from your spoken summary.

Limitations and responsible use

AI-generated clinical notes have real limitations, and being straightforward about them matters more than overpromising.

• Content accuracy is not guaranteed. AI models can occasionally generate content not present in the original conversation, omit details that were discussed, or place information in the wrong section. Every note requires clinician review.
• Complex encounters are harder. Visits with multiple active problems, frequent topic switches, or extensive back-and-forth between clinician and patient can produce drafts that need more editing than straightforward encounters.
• Medical terminology varies. Regional terminology, abbreviations specific to your practice, and newer drug names may not always be captured correctly. You can correct these during review, and the system improves over time with use.
• The AI does not make clinical decisions. The Assessment and Plan sections reflect what was discussed during the encounter, not independent diagnostic or treatment recommendations from the AI. Clinical judgment remains entirely with the clinician.

These limitations are why Dictum positions its output as review-ready documentation rather than finished notes. The tool helps reduce documentation burden — it does not replace the clinician’s responsibility to verify what goes into the patient’s chart. See Dictum pricing plans for details on available features and usage limits.