Reviewing AI Suggestions
The AI assistant drafts content for your review. You decide what to send, edit, or discard. Every AI action is logged and under your control.
DPC Pro’s AI assistant helps clinicians by drafting message replies, answering practice questions, and compiling daily summaries. This page explains how AI-generated content appears in your workflow, how to review it effectively, and how to provide feedback that improves future suggestions.
The AI is a support tool. It does not make clinical decisions, send messages to patients, or modify records without your explicit approval.
AI Transparency
Section titled “AI Transparency”Understanding what the AI does and does not do helps you use it with confidence.
What the AI Does
Section titled “What the AI Does”| Capability | How It Works |
|---|---|
| Drafts message replies | Reads the patient’s message, considers their clinical history and your practice’s communication style, and suggests a response |
| Answers practice questions | Retrieves data from your patient panel, schedule, and billing to answer questions you ask in the chat |
| Compiles daily summaries | Gathers your schedule, unread messages, and patient panel changes into a morning briefing |
What You Control
Section titled “What You Control”| Area | Your Control |
|---|---|
| Message drafts | You review, edit, and send (or discard) every draft before it reaches a patient |
| Practice questions | You decide what to ask and whether to act on the answer |
| Daily summaries | You choose which sections to receive and when they are delivered |
| Guardrails | Your practice manager configures what the AI can access and do. You can request changes to these settings. |
What the AI Does Not Do
Section titled “What the AI Does Not Do”- Make diagnoses: the AI does not provide definitive clinical diagnoses. It can reference conditions in a patient’s record but does not determine what a patient has.
- Prescribe treatment: the AI does not recommend medications, dosage changes, or treatment plans.
- Contact patients directly: no message, email, or text reaches a patient without a clinician or staff member sending it.
- Modify records: the AI can read patient records but cannot write to clinical notes, prescriptions, or billing records.
- Access external sources: the AI has no internet access and cannot reach external databases, drug references, or medical literature.
Audit Trail
Section titled “Audit Trail”Every AI interaction is logged:
- Who initiated the action (your user account)
- What data was accessed (patient records, schedule, billing)
- When the action occurred
- What type of action it was (conversation, data retrieval, draft generation)
You can review your own AI audit entries at Compliance → Audit Log. Practice managers see the full audit log across all users. See Review AI Actions for details.
Reviewing AI-Drafted Message Replies
Section titled “Reviewing AI-Drafted Message Replies”When a patient sends a message, the AI may draft a suggested reply based on:
- The content of the patient’s message
- The patient’s clinical history (allergies, conditions, medications, recent notes)
- Your practice’s communication patterns
How Drafts Appear
Section titled “How Drafts Appear”AI-drafted replies appear in the messaging interface when you open a patient conversation. The draft is clearly labeled as AI-generated and is not visible to the patient.
Your Review Workflow
Section titled “Your Review Workflow”- Read the patient’s message: understand what the patient is asking before looking at the draft.
- Review the draft: check it for accuracy, tone, and completeness.
- Edit as needed: modify the text to match your clinical judgment and personal communication style.
- Send: once you are satisfied, send the message. It goes to the patient as a message from you.
- Or discard: if the draft is not useful, discard it and write your own response.
When to Write From Scratch
Section titled “When to Write From Scratch”AI drafts work well for routine responses: refill confirmations, visit follow-ups, and common health questions. Write your own response for:
- Complex clinical discussions that require nuanced explanation
- Sensitive topics (mental health, end-of-life, difficult diagnoses)
- Situations where the AI draft misunderstands the patient’s question
- Any response where you are not confident in the draft’s accuracy
For the full messaging reference, see Send and Receive Messages and AI-Drafted Message Replies.
Using the AI Chat
Section titled “Using the AI Chat”Beyond message drafts, you can interact with the AI through the chat interface.
Asking Practice Questions
Section titled “Asking Practice Questions”Open the AI chat and ask questions like:
- “What is my schedule for tomorrow?”
- “Which patients have diabetes on my panel?”
- “What medications is [patient name] currently taking?”
- “How many unread messages do I have?”
The AI retrieves data from your practice database and presents it in the conversation. Each data retrieval is logged in the audit trail.
Reviewing Answers
Section titled “Reviewing Answers”When the AI provides an answer:
- Verify key facts: for clinical data, confirm against the patient record if the information is driving a care decision.
- Ask follow-ups: the AI maintains conversation context, so you can drill deeper.
- Act on your judgment: the AI surfaces information; clinical decisions are yours.
See How the AI Works for the full overview of AI capabilities and data access.
Reviewing Daily Summaries
Section titled “Reviewing Daily Summaries”The daily summary is delivered to your dashboard each morning (and optionally by email). Your clinician summary includes:
- Today’s schedule: your patient visits for the day
- Unread messages: conversations assigned to you
- Patient panel changes: new patients, membership changes
The summary is informational. It compiles data from your practice database and presents it. It does not make recommendations or prioritize patients by urgency.
To customize your summary:
- Navigate to Settings → AI Assistant → Daily Summary.
- Toggle sections on or off.
- Change delivery time or disable summaries entirely.
See Daily Practice Summaries for configuration details.
Adjusting AI Behavior
Section titled “Adjusting AI Behavior”Providing Feedback on Drafts
Section titled “Providing Feedback on Drafts”The most effective way to improve AI suggestions is through consistent editing:
- If drafts are too formal, edit them to a warmer tone repeatedly.
- If drafts miss important context, add the missing information each time.
- Over time, the AI adapts to the patterns in your edits.
Requesting Guardrail Changes
Section titled “Requesting Guardrail Changes”If you want the AI to have more (or less) access to patient data, or if you want features enabled or disabled, talk to your practice manager. Guardrail settings are managed at the practice level.
See AI Guardrails and Preferences for the configuration reference.
Disabling AI Features
Section titled “Disabling AI Features”If you prefer not to use AI-drafted replies or the AI chat:
- Ask your practice manager to adjust your preferences.
- You can always ignore AI drafts and write your own messages.
Related Pages
Section titled “Related Pages”- Quick Start for Clinicians: getting started
- Daily Clinical Workflow: your daily routine
- How the AI Works: full AI overview
- AI-Drafted Message Replies: draft reply details
- AI Guardrails and Preferences: configuration and controls
- Review AI Actions: audit trail for AI interactions
- Daily Practice Summaries: morning briefing configuration
Need Help?
Section titled “Need Help?”If you have questions about AI suggestions, reach out to the DPC Pro support team at [email protected] or visit the troubleshooting guide.