AI Scribe in Dermatology and Procedure-Heavy Specialties

The procedure-visit mix

Dermatology, ophthalmology, urology, gastroenterology, and several other specialties share a documentation pattern: most days mix office visits with brief in-office procedures. The visit volume is high, the documentation per visit is moderate, and the procedure documentation has its own structure that office-visit AI scribes don't naturally handle.

This mix makes AI scribe adoption in these specialties more nuanced than in general primary care. The technology helps with the office-visit portion. The procedure documentation often needs to stay manual or use specialty-specific templates. Getting the workflow right requires thinking about both layers.

This article covers how procedure-heavy specialties should think about AI scribes: where they help, where they don't, and how to set up a workflow that gets the benefit without the friction.

What works well: the consultation portion

The conversational portion of a specialty visit (history-taking, discussion of findings, treatment options, plan) maps cleanly to what AI scribes are designed for. The accuracy and time savings on this portion are comparable to primary care.

For example, in a dermatology consultation:

• The patient describes the lesion or concern
• The dermatologist asks targeted questions
• The dermatologist explains the diagnostic and treatment options
• The patient asks questions; the dermatologist answers
• The plan is agreed

All of this is conversation. AI scribes capture it well. The note's HPI, assessment, discussion-of-options, and plan sections fill in naturally.

What doesn't work as well: the structured procedure documentation

The lesion description, the operative-style narration of a procedure, and the post-procedure instructions often have specific structural requirements that AI scribes don't generate cleanly.

In dermatology specifically:

Lesion descriptions need anatomic precision. "1.2 cm pearly papule with central ulceration and telangiectasias, located on the left preauricular cheek." The AI captures the words, but the exact anatomic position needs to come from a structured location selector in the EHR, not from prose.

Multi-lesion visits multiply this. A full-body skin check identifying 12 lesions for biopsy or treatment doesn't map well to ambient documentation. Each lesion needs its own structured entry.

Procedure notes need their own format. A shave biopsy, a punch biopsy, an electrodessication and curettage, a cryotherapy treatment, an excision. Each has a standard documentation format with specific elements (consent, anesthesia, size, technique, hemostasis, closure, complications, pathology sent). AI scribes don't generate these in their standard ambient mode.

In ophthalmology:

Slit lamp findings have their own structured format. AI scribes can capture verbal description but the structured exam findings (intraocular pressures, visual acuities, refraction) come from instruments, not conversation.

Procedure documentation for intravitreal injections, laser treatments, and minor procedures has specific elements. Date, eye, drug, lot, technique, complications. Often structured in the EHR.

In urology:

Cystoscopy findings and PSA tracking have structured patterns. The conversational discussion of findings can be ambient; the structured findings come from the procedure itself.

Procedure notes for in-office procedures (vasectomy, biopsy, etc.) have standard formats.

The pattern across these specialties: AI scribes handle the office-visit conversation well; structured procedure documentation needs its own workflow.

Workflow patterns that work

A few patterns specialty practices have settled on:

The hybrid note pattern. The AI scribe generates the visit narrative; the EHR's procedure documentation form is filled out separately (often by the provider or a structured-data entry assistant). The final chart has both: a prose note from the AI and a structured procedure documentation entry.

The dictation handoff pattern. During an in-office procedure, the provider switches from ambient AI scribe to traditional dictation for the procedure narrative. The procedure-specific dictation goes into a procedure note template. The visit note (with the conversation before and after the procedure) uses the AI scribe.

The macro template pattern. For high-volume repetitive procedures (e.g., a dermatologist doing 8 cryotherapy treatments per day), a structured EHR macro generates 80% of the procedure note from a few clicks; the AI scribe captures the consultation portion. The combination is fast.

The MA documentation pattern. Some specialty practices have an MA document the procedure structured fields while the provider does the procedure. The AI scribe handles the consultation. The combined documentation is comprehensive.

Each pattern has tradeoffs. Practices typically find their own variation after a few months of experimentation.

What to do during the procedure itself

A practical question: when you're performing a procedure, is the AI scribe still recording? Should it be?

The answer depends:

Yes, leave it on. If the procedure is brief, doesn't involve significant verbal narration, and the audio is mostly clinical conversation (explaining what you're doing to the patient), leaving the AI on captures useful content.

Pause it. If the procedure involves prolonged silence, music in the room, or sensitive content (e.g., a dermatology biopsy where the patient is anxious and you're focused on technique), pausing during the procedure and resuming after may produce cleaner notes.

Switch to dictation. If you're trained on dictation and the procedure has a standard narrative, switching to dictation for the procedure portion is often faster than expecting AI scribe to handle it.

The choice is workflow-dependent. Many specialty practices try all three patterns in different visit types and settle on what works.

Vendor questions specific to procedure specialties

If you're in a procedure-heavy specialty, the vendor evaluation should include:

1. Does your platform support specialty-specific templates for [my specialty]?
2. Can the platform pause and resume mid-visit without losing context?
3. How does the platform handle lesion descriptions with specific anatomic locations?
4. Does the platform integrate with my EHR's procedure documentation forms?
5. Do you have customers in [my specialty]? Can I talk to one?
6. How does your platform handle visits with multiple procedures or multiple lesions?

The vendors with specialty experience have specific answers. Generic answers usually mean they haven't deeply integrated for your specialty.

The time savings, honestly

In procedure-heavy specialties, AI scribe time savings are typically 50 to 75% of what primary care sees. The reason: a meaningful portion of the documentation work is procedure-related, and the AI doesn't reduce that work much.

For a typical dermatology practice:

• AI scribe saves 15 to 25 minutes per day on consultation documentation
• Procedure documentation takes its usual time (the AI doesn't help much)
• Total daily time recovered: 20 to 40 minutes

For a typical ophthalmology practice:

• Similar pattern: AI scribe helps on the consultation portion, procedure documentation continues to use existing workflows
• Total daily time recovered: 15 to 35 minutes

The time savings are still net positive and usually justify the subscription cost. They're just not the dramatic time savings primary care providers report.

The accuracy considerations

Specialty practices have specialty-specific terminology that AI scribes may not have been trained extensively on:

Dermatology vocabulary. Specific lesion morphology terms (verrucous, lichenified, telangiectatic), specific dermatologic conditions (lichen planus pemphigoides, granuloma annulare). The AI's accuracy on common terms is fine; on rare terms, it's variable.

Ophthalmology vocabulary. Eye-specific terminology (proptosis, exophthalmos, hyphema, intraocular pressure terminology). Some platforms handle this well; some don't.

Surgical specialty vocabulary. Procedure names, anatomic landmarks, instrument names. The AI captures common terms well; uncommon terms may be misheard.

The mitigation: customize the AI's vocabulary with your specialty's specific terminology. Most platforms support adding custom vocabulary. The 30 minutes you invest in this pays back in reduced editing time.

The case for adoption in procedure-heavy specialties

Despite the more complicated picture, AI scribe adoption in procedure-heavy specialties is usually positive. The reasons:

• The consultation portion of visits is real work, and AI scribes do reduce its documentation burden
• The time savings, while smaller, are still meaningful (30 to 60 minutes per day for a busy specialist)
• Patient satisfaction improves; the provider looks more engaged during the conversational portion of visits
• The technology will continue to improve, and specialty-specific features are being added regularly

For dermatology, ophthalmology, urology, GI, and similar specialties, adopting an AI scribe in 2026 is a reasonable choice. Expect modest time savings, plan for hybrid documentation, and choose a vendor with at least some specialty experience.

The case for waiting in procedure-heavy specialties

Some specialty contexts where waiting another year or two might make sense:

High-volume procedure-only practices. A practice that does almost entirely procedures with minimal consultation (some dermatology MOHS surgery practices, some pure-procedural urology practices) gets less benefit. The procedure documentation is the main work, and AI scribes don't address it.

Practices with strong existing templating systems. A practice that has already invested heavily in EHR macro templates for their specialty may find the marginal benefit of AI scribes smaller. Their template system is already capturing what AI scribes would capture.

Practices with vendors that don't yet serve your specialty well. If the major AI scribe vendors don't have strong specialty support for your area, waiting until the specialty-specific tools mature is reasonable.

For practices that fit any of these, evaluating again in 12 to 18 months is sensible. The vendors are improving and the specialty coverage is expanding.

The dermatology MOHS surgery case specifically

Worth a brief mention because it's a frequent question. Mohs surgery practices have intensive procedure documentation (each stage, each margin, each tissue piece). AI scribes don't help with this. The consultation portion (deciding to proceed, post-op counseling) is captured well by AI scribes but is a small fraction of the documentation burden.

Most MOHS practices that have evaluated AI scribes conclude that the cost-benefit doesn't justify it for the Mohs portion of the practice. For Mohs surgeons who also do general derm clinic on certain days, the AI scribe can help on the clinic days but not on the Mohs days. That split is workable.

---

For the broader specialty considerations, see the AI scribe vs medical assistant cost comparison which addresses staffing patterns common in specialty practices. For consent considerations across specialties, talking to patients about AI scribes covers the script that works across visit types.