Skip to main content

Voice-Powered Prescription Creation

Aria’s Smart Medical Dictation allows you to create prescriptions by speaking naturally. Our AI automatically identifies and highlights medications, dosages, symptoms, and medical terminology in real-time.
No need to follow rigid formats or templates - just speak as you normally would during patient consultation.

How It Works

1

Start dictating

Click the microphone icon and speak naturally about the patient’s condition and treatment
2

Real-time entity recognition

AI identifies and highlights:
  • Medications: Automatically detected and styled with medication formatting
  • Symptoms: Highlighted with visual indicators
  • Dosages: Extracted with proper units (mg, g, ml, etc.)
  • Medical terminology: Recognized and validated
3

Review and confirm

Check the detected entities shown as badges below the dictation box
4

Finalize prescription

Detected medications are automatically added to the prescription with proper dosages

Example Dictation Flow

Here’s what happens when you dictate a typical prescription:
"Patient complains of fever and headache for 3 days. 
Prescribing Paracetamol 500mg and Cetirizine 10mg for symptom relief."

Visual Entity Recognition

As you dictate, entities are highlighted in different styles:

Medications

Highlighted with subtle background and borderExample: Paracetamol 500mg

Symptoms

Highlighted with amber background for quick identificationExample: fever, headache
The visual highlighting helps you quickly verify that the AI correctly understood your dictation before finalizing the prescription.

Detected Entity Badges

Below the dictation text, you’ll see real-time badges showing detected entities:
Detected:
  • 💊 Paracetamol 500mg (Medication)
  • 💊 Cetirizine 10mg (Medication)
  • ⚡ Fever (Symptom)
  • ⚡ Headache (Symptom)

Supported Medical Entities

The AI recognizes a wide range of medical entities:

Medications

  • Generic drug names (Paracetamol, Metformin, etc.)
  • Brand names (Crocin, Dolo, etc.)
  • Dosage forms (tablets, capsules, syrup, injections)
  • Dosage strengths with units (500mg, 1g, 10ml)

Symptoms & Conditions

  • Common symptoms (fever, cough, headache, nausea)
  • Chronic conditions (hypertension, diabetes)
  • Acute conditions (URI, gastritis, migraine)

Medical Terminology

  • Vital signs (BP, temperature, pulse)
  • Laboratory values (blood sugar, HbA1c)
  • Clinical observations
The entity recognition engine is trained on Indian medical terminology and commonly prescribed medications in India, making it highly accurate for local practice patterns.

Best Practices

For Accurate Recognition

1

Speak clearly

Use a normal speaking pace - not too fast, not too slow
2

Include dosages

Always mention dosage amounts with medications (“Paracetamol 500mg” not just “Paracetamol”)
3

Use standard terminology

Prefer generic names over colloquialisms when possible
4

Verify detected entities

Always check the badge display before finalizing

Common Dictation Patterns

Patient presents with acute upper respiratory infection.
Fever 99.8°F, cough, sore throat.
Prescribing Azithromycin 500mg OD for 3 days,
Paracetamol 650mg SOS for fever.
When the AI detects a medication name during dictation, it automatically links to the Smart Medication Search database:
If you say “Crocin”, the system will offer you specific dosage options from the ABHA-compliant medication database (Crocin 500mg, Crocin 650mg, Crocin Advance, etc.)
  1. You dictate a partial or complete medication name
  2. AI matches it against 100,000+ medications in the database
  3. You get instant suggestions with:
    • Available dosage strengths
    • Composition details
    • Common administration routes
  4. Select the exact medication variant with one click

Dictation Interface

The dictation interface provides clear visual feedback:

Active Dictation Indicator

  • Pulsing microphone icon shows recording is active
  • Blinking cursor indicates where text is being added

Live Text Display

  • Text appears in real-time as you speak
  • Entity highlighting updates dynamically
  • Scroll automatically follows new text

Entity Badge Panel

  • Updates live as entities are detected
  • Color-coded by type (medications, symptoms)
  • Click badges to edit if needed
Always review the final prescription before sending to the patient. While the AI is highly accurate, medical documentation requires human verification.

Error Handling

Unrecognized Terms

If the AI doesn’t recognize a medication or term:
1

Check spelling

The term appears without highlighting - verify you spoke it correctly
2

Use medication search

Manually search for the medication using Smart Medication Search
3

Add manually

You can always type or select medications manually if dictation doesn’t catch them

Incorrect Detection

  • Click on the detected entity badge
  • Choose “Remove” or “Edit”
  • The text highlighting will update accordingly
  • You can re-dictate that portion or type it manually

Performance Tips

Environment

Use in a quiet environment for best recognition accuracy. Background noise can affect entity detection.

Microphone

Use a good quality microphone or headset. Built-in laptop mics work but dedicated mics are better.

Internet

Stable internet connection required for real-time AI processing and entity recognition.

Language

Currently optimized for English medical terminology. Hindi and regional language support coming soon.

Next Steps

Smart Medication Search

Learn how to quickly search and add medications from the 100,000+ drug database

Complete Prescriptions

Understand the full prescription creation workflow

Patient History

Access complete patient timelines while creating prescriptions
Pro Tip: Combine dictation with the Patient Health Assistant. Ask “Any drug allergies?” via voice before prescribing to get instant answers from patient history.

Build docs developers (and LLMs) love