Stop
Stop using the AI output and do not send related patient-facing, payer-facing, or internal decision content.
AI incident response guide
Incident escalation after wrong-tool dental AI use
When patient information may have entered the wrong AI tool, the safest response is not panic or deletion. The practice needs a calm escalation path: stop use, preserve facts, notify the right owner, document what happened, and follow the practice's incident process.
Best for: Dental owners, office managers, privacy leads, billing teams, front desk teams, and clinical staff who need a clear wrong-tool AI escalation path.
Asset promise: Practice-owned workflow asset. Use this resource to create one reviewed script, checklist, SOP, packet, scorecard, or decision map the practice can save and reuse.
Create an incident pathSee all resources
Sensitive patient, payer, claim, balance, screenshot, and schedule context belongs in approved practice-owned workflows. Use Smarter Practice AI for practice-owned workflows, then let the right team member finish the output before it is used.
The first-hour rule
A small mistake becomes harder to evaluate when staff delete evidence, stay quiet, keep using the output, or try to fix the issue alone. The practice should give staff a non-punitive reporting path that preserves facts and routes the issue to the privacy lead or owner.
Preserve facts
Record what was entered, which tool was used, when it happened, who used it, and whether any output was saved or sent.
Notify
Notify the office manager, owner, privacy lead, or designated incident owner immediately.
Review
Follow the practice's incident process, including risk analysis, vendor review, corrective action, and any required next steps.
What to document
- Tool or website name and account type.
- Whether the tool was approved or personal/unapproved.
- What information was pasted, uploaded, screenshotted, transcribed, or generated.
- Whether the input included names, appointment details, symptoms, payer data, balances, screenshots, EOBs, images, audio, or messages.
- Whether output was copied, saved, sent, printed, or used for a decision.
- Who was notified and when.
- Immediate containment steps and follow-up owner.
Source note: HHS Breach Notification Rule HHS Security Rule risk analysis guidance HHS Security Rule HHS OCR enforcement process
Staff huddle script
Manager script
If you think patient, payer, EOB, screenshot, audio, or message details went into the wrong AI tool, stop and tell me immediately. Do not delete everything and hope it goes away. We need the facts so the practice can review the issue, protect patients, and improve the workflow.
Non-punitive framing
The fastest safe path is early reporting. The goal is to correct the workflow, preserve facts, and avoid making a small issue larger.
Policy language to adapt
Wrong-tool AI escalation rule
Any staff member who may have entered patient, payer, claim, schedule, image, audio, screenshot, or message information into an unapproved AI tool must stop using the output, notify the designated manager or privacy lead immediately, and document the facts needed for review.
No deletion-only rule
Deleting a chat, file, transcript, screenshot, or output is not a substitute for reporting. Staff should preserve enough information for the practice to assess what happened and follow its incident process.