Dictation vs AI: The Best Way to Write Clinic Letters in 2026
Compare dictation, speech recognition and AI drafting by input methods, clinical templates, review and the wider clinic workflow.
Dr. Shady Nafie
Consultant Urologist & Founder of Docyment ·
Last updated: July 2026
I dictated my clinic letters for eighteen years.
Handheld Philips recorder. End of every clinic, or sometimes on the train home. Secretary picks up the files, types them the next morning, sends them to me for review by afternoon. Sometimes same-day if she wasn't buried. Sometimes two days later.
It worked. Sort of. I knew the letters would get done eventually. I also knew that somewhere between dictation and the GP's inbox, my patient was waiting for information their GP needed.
The question I get from colleagues is whether to change the capture and drafting workflow. This article compares approaches to clinic letters so you can test what fits your practice.
What Are the Four Methods for Producing Clinic Letters in 2026?
This article considers secretary-supported dictation, speech recognition, consultation-recording tools and drafting from clinician-entered findings. Some products combine several methods. Compare current capabilities and test the complete review workflow.
Before comparing, it helps to be clear about what we're actually comparing. There are four main approaches consultants use to produce clinic letters in UK practice:
- Traditional dictation → secretary typing
- Voice recognition software (Dragon Medical)
- Ambient AI scribes (recording-based)
- Structured AI letter generation (input-based)
Each changes the division of work between capture, drafting, correction and approval. Compare how it performs in your own clinic.
Method 1: Traditional Dictation → Secretary
In a secretary-supported workflow, you dictate and your secretary prepares the typed draft.
How it works: You use a handheld recorder, a dictation app, or a phone to record each letter after the consultation. The audio file goes to your secretary, who transcribes and formats it. You receive a draft, correct any transcription errors, and sign it off.
Turnaround and cost: Check the current service terms and measure the complete capture, correction and approval workflow.
What to test:
- How well the draft matches your clinical terminology and templates
- How corrections and approval fit your workflow
What needs attention:
- Check omissions, terminology, recipient details and clinical decisions
- Tell the patient and obtain consent if recording the consultation
- Review data processing and retention arrangements
The hidden cost: Delegating typing does not remove your input and review. Measure those tasks when comparing another workflow.
Method 2: Medical Dictation Software UK — Dragon Medical
Speech recognition turns spoken words into text. Test the exact product version and template workflow you are considering.
How it works: Speak into a microphone, then review the transcription in a document or template.
Turnaround and cost: Check the current service terms and measure the complete capture, correction and approval workflow.
What to test:
- The quality of the captured source
- Whether the output matches your clinical formats
- The correction and approval workflow
What needs attention:
- Check omissions, terminology, recipient details and clinical decisions
- Tell the patient and obtain consent if recording the consultation
- Review data processing and retention arrangements
The honest take: Transcription is one part of producing a clinical document. Judge the full workflow rather than the speed of text appearing on screen.
Method 3: Ambient AI Clinic Letters (Recording-Based)
Consultation-recording tools can draft from a transcript. Check the particular product's input methods, output formats and review controls.
How it works: A device captures the consultation for transcription and drafting. Check the draft against the source and approve it before use.
Turnaround and cost: Check the current service terms and measure the complete capture, correction and approval workflow.
What to test:
- The quality of the captured source
- Whether the output matches your clinical formats
- The correction and approval workflow
What needs attention:
- Check omissions, terminology, recipient details and clinical decisions
- Tell the patient and obtain consent if recording the consultation
- Review data processing and retention arrangements
The honest take on consultation recording: Test it with the kind of consultation you conduct and the document you need. A category label does not establish the editing required or suitability for your clinic.
Method 4: Structured AI — Enter Findings, Get a Formatted Letter
This is the approach I use now, and the one that prompted me to build Docyment.
How it works: After each consultation, you give the AI your clinical content — the method is your choice. Docyment has four input modes:
- Dictate — speak your clinical findings
- Record — capture the consultation with the patient's consent
- Write — type or paste your findings
- Scan — photograph a note, letter or result
The selected template guides the draft. An AI reviewer checks it, and you review and approve it. Clinical documents are currently produced in English.
Turnaround and cost: Check the current service terms and measure the complete capture, correction and approval workflow.
What the workflow offers:
- Dictate, Record, Write and Scan
- One consultation can produce a clinic letter, referral and prescription
- Drafting in your clinical templates, followed by review and approval
What needs attention:
- Check omissions, terminology, recipient details and clinical decisions
- Tell the patient and obtain consent if recording the consultation
- Review data processing and retention arrangements
The honest take: Choose the capture method that fits the consultation. Docyment connects document production with bookings and patient records, so consider the wider clinic workflow as well as the draft.
Direct Comparison: Which Method Wins?
| Evaluate | Ask in your trial |
|---|---|
| Capture | Does the input method fit your consultations? |
| Format | Can it produce your clinical templates? |
| Review | Can you compare the draft with its source and correct it? |
| Team | How does it fit booking, patient records and approval? |
| Terms | Are processing arrangements, features and fees clear? |
Real Clinical Scenarios: How Each Method Performs
Post-operative urology review (PSA result, MRI findings, decision on treatment pathway)
This is a complex letter — multiple data points, a nuanced management decision, potential for the GP to forward it to the patient.
For each method: Check patient identification, clinical findings, decisions and the complete review workflow. Test the templates your clinic needs.
Post-paediatric ENT clinic (multiple children, straightforward grommets follow-ups)
High volume, lower complexity per patient. Efficiency matters most.
For each method: Test the complete capture and review workflow across the consultations you conduct. High volume does not remove the need to check each document.
What I Actually Recommend
After sixteen years of traditional dictation, two years of Dragon Medical, several months of evaluating ambient AI scribes, and now daily use of Docyment:
If you have a reliable, experienced secretary and your current turnaround is acceptable: Traditional dictation works. Don't fix what isn't broken — but know that your dependency on one person is a single point of failure.
If your secretary has left or you are setting up a practice: Compare capture and review options before committing. Test your templates and measure the administrative work.
If you work in NHS primary care or emergency care: Check which products your organisation approves and how they fit its clinical-record workflow.
If you are a private practice consultant producing outpatient letters: Test clinician-entered findings alongside the other capture methods. Try Docyment free to see how it fits your clinic.
The goal isn't the most sophisticated technology. It's letters to GPs on the day they should arrive, with content you trust, produced in time you can afford to spend.
The useful question is which workflow helps your team complete documents with the review they need.
Frequently Asked Questions
Is AI faster than dictation for clinic letters?
It depends on the consultation and review required. Measure capture, correction and approval with comparable documents during a trial.
Should I replace Dragon Medical with AI?
Compare the exact workflows you use. Docyment accepts Dictate, Record, Write and Scan and drafts into your chosen template. Check whether it changes the work you need to do.
Do recording-based tools produce UK clinic letters? Output formats vary by product and plan. Ask for a trial with your letter templates instead of assuming a recording-based tool only produces notes.
How should I compare the cost of clinic-letter tools? Check subscription, staffing and service costs alongside review effort. Docyment Free is for one person with 20 document generations a month. UK Pro is £20 per clinician seat a month; admin and reception seats are free and unlimited on Pro. See current pricing.
Related Reading
- Clinical Letter Software for UK Consultants: What's Actually Worth Using — the full landscape beyond AI
- AI Medical Letters UK: What Consultants Need to Know — ambient recording vs structured input compared in detail
- Why I Built an AI Clinical Letter Generator — the origin story and what Docyment does and doesn't do
Try Docyment free — no credit card required, and Free is for one person, with 20 document generations a month.