AI Assistant Guide
How to use AI features for faster, more accurate radiology reporting
Enter your findings and let the AI produce a structured radiology report with Technique, Findings, and Impression sections.
- Enter findings in the Relevant Findings box (type or dictate)
- Click Generate Full Report or press Ctrl+Alt+G
- Review and edit the generated report
- Click Accept and Sign or press Ctrl+Alt+Enter
Example:
Input: "CT chest with contrast. 45yo male, chest pain. No PE. Mild cardiomegaly. Small bilateral effusions."
Output: Full structured report with PROCEDURE, TECHNIQUE, FINDINGS, and IMPRESSION sections formatted to radiology standards.
AI can be incorrect. Always review generated reports for accuracy before signing. AI is a tool, not a replacement for clinical judgment.
The AI doesn't just format — it improves your input. You can be messy, use abbreviations, or speak casually.
Typo correction
In: "pulmonry embolizm"
Out: "pulmonary embolism"
Terminology upgrade
In: "big heart, fluid in chest"
Out: "cardiomegaly, pleural effusions"
Abbreviation expansion
In: "SOB, no PE"
Out: "shortness of breath, no pulmonary embolism"
Thought completion
In: "CT shows... some fluid"
Out: "CT demonstrates bilateral pleural effusions"
Relevant Findings has two in-browser options (same Google Cloud Speech setup):
- Medical dictation (Google) — medical dictation model for clinical and radiology wording. Record, tap stop, then transcription appears.
- Live dictation — Google (general) — general Speech-to-Text; text arrives in chunks while you speak (not the medical model).
- Choose Medical dictation (Google) or Live dictation — Google (general)
- Speak into the microphone; stop when finished (Medical) or tap stop when done (both)
- Text appears in the Relevant Findings box (live mode updates as chunks return)
Voice commands: Say "period", "comma", "next line", "next paragraph", "impression", etc. Medical dictation is best for these. The AI still cleans up formatting when it generates the report.
Report Content editor: There is no voice dictation button next to the main report body. Edit there by keyboard, paste, or your external dictation workflow. Voice for AI input is only under Relevant Findings.
Shortcut: With focus in Relevant Findings, F9 toggles Medical dictation (Google) (not the general live button).
Preset finding buttons appear below the Relevant Findings box, tailored to the current exam type. One tap adds the text to your findings.
Click the gear icon to add, remove, or reorder findings for each exam type.
Helpful Features
Copy to Findings
Special Instructions from the order can be copied to Relevant Findings with one click.
Document Communication
Note verbal discussions with the referring physician directly in the report.
Prior Reports
If prior reports exist for this patient, they appear at the top of the page for reference.
Templates
Load pre-built report templates for consistent formatting. Expand Template Options to access.
Best Practices
Writing Effective Findings
- Be specific and detailed
- Include clinical context
- Note comparison with priors
- Include both positive and negative findings
Quality Assurance
- Always review AI output
- Verify medical accuracy
- Check findings match impression
- Include relevant measurements
Voice Dictation Tips
- Speak clearly at normal pace
- Use medical terms confidently
- Minimize background noise
- Review transcription before generating
Keyboard Shortcuts
- Ctrl+Alt+G — Generate Report
- Ctrl+Alt+Enter — Accept and Sign
- F9 — Medical dictation (Google) when Relevant Findings is focused
Important Guidelines
- AI-generated content must always be reviewed for accuracy
- AI is a tool — it does not replace clinical expertise
- Ensure reports meet quality, safety, and regulatory standards