Patient portal AI assistant
Epic Emmie (incl. Ask Emmie)
Epic Emmie combines direct patient record questions with health-system control of available actions. Sutter’s published version cannot schedule, message, or update the chart. Its chat-visibility rules do not establish an absence of retention or secondary processing.
Checklist assessment dated October 30, 2026. Independent confirmation is incomplete. Hugo has not personally verified this service.
Agency checklist · October 30, 2026
Leans constrains 4 of 12 yes
Document-based draft. Independent confirmation is incomplete. Hugo has not personally verified this service.
Assessed function: AI assistant (incl. Ask Emmie) inside the MyChart patient portal, deployed by health systems
A no can mean missing documentation or institution-dependent control; it does not by itself establish harmful behavior. These categories are not clinical safety or product-quality ratings.
- Public documents do not establish actual product behavior.
- The independent review screened rules and selected evidence; it did not replicate all twelve answers for this tool.
See all twelve answers and their evidence
| Question | Answer | Evidence / review | Source and reasoning |
|---|---|---|---|
| 1. Told it is AI | No | deployer-dependent Not independently replicated | Emmie uses generative AI, a technology that can sometimes make mistakes.Source Epic's page calls Emmie "Epic's built-in AI assistant," and Sutter and Reid both label it AI in their help pages. I found no Epic document saying the in-product experience always shows an AI notice, so this rests on each health system's pages. Reviewer may reasonably flip this. |
| 2. Operator and model maker named | No | not-documented Not independently replicated | Emmie was developed by EpicSource Epic is named as developer and the health system as operator. No page I read names the maker of the underlying AI model. Epic's Emmie page does not, and neither do the Sutter, Reid or TechTarget pages. |
| 3. Patient chooses to use it | Yes | documented Not independently replicated | enable the Preview Features toggleSource At Reid, patients turn Emmie on in their own MyChart settings. At Sutter it appears on the homepage after sign-in and the patient opens it. The chat starts only when the patient uses it. Other Emmie features, such as AI result summaries, are shown without the patient asking (see Q5). The toggle is a rollout choice by the health system. |
| 4. Patient sets the task | No | documented Not independently replicated | saving 750 hours of staff timeSource Borderline. The patient chooses the questions. The operator limits Emmie to portal and chart topics and has a separate Billing Emmie. Epic markets it by staff-time and call-volume savings, which are operator goals. Sutter lists what it cannot do, such as scheduling or messaging the care team. |
| 5. Patient can decline or stop | Yes | Self-started rule; direct refusal wording not documented Not independently replicated | Emmie is designed to help patients find information and navigate My Health Online more easily, not replace care teams or support staff.Source Yes follows the adopted Q5 rule for a self-started optional tool, rather than an explicit published refusal right. The original coder's note below predates that normalization and is retained as history: No source states that a patient can decline or stop Emmie and keep the same service. Not using the chat appears possible and human support stays available, which is my inference. The Sutter page metadata mentions "how to opt in or out" but the FAQ text has no opt-out answer. Epic's page shows AI result summaries and reminders appearing without a patient opt-in. |
| 6. Patient sees the AI output | Yes | documented Not independently replicated | AI-generated result summaries that appear as soon as the result is available in MyChartSource The patient sees Emmie's own answers and AI summaries directly in MyChart, not a later human rewrite. |
| 7. Patient can correct or challenge | Yes | documented Not independently replicated | to provide feedback in EmmieSource Weak route. Sutter's answer to "What should I do if I disagree with an answer Emmie gives me?" is thumbs-up/thumbs-down feedback plus following up with the care team. Nothing promises a correction or review. Reid says patients can "request updates to allergies, medications, and health issues," which corrects the chart, not Emmie's output. Reviewer may reasonably flip this. |
| 8. Patient can get a copy | No | not-documented Not independently replicated | Emmie chat logs aren’t stored as part of your legal medical record and your provider won’t see them.Source No export or copy of chats is documented. Chat logs are outside the legal medical record, so the normal record-access route may not reach them. |
| 9. Patient can delete | No | not-documented Not independently replicated | Quotation not recorded; evidence repair pending. Source No way for patients to delete Emmie chat history is documented in the sources I read. |
| 10. Training or sale does not override patient choice | No | not-documented Not independently replicated | Quotation not recorded; evidence repair pending. Source No source states whether chat or chart data is used to train models, or offers an opt-out. Reid says only that "no health information is shared outside the patient's established care relationship." That does not address training. |
| 11. Patient can run it or take it elsewhere | No | not-documented Not independently replicated | There’s no separate app or download.Source Emmie exists only inside a health system's MyChart. Nothing documents a way for the patient to run it or move it elsewhere. Epic's own page returned 403 to direct fetch, so I read an Internet Archive copy of https://www.epic.com/software/emmie/, which lists features only. |
| 12. No advertising or data business | No | deployer-dependent Not independently replicated | Emmie operates within My Health Online and follows the same privacy and security protections as the rest of the platform.Source Data handling is set by each deploying health system. The Sutter FAQ only points to the platform's general protections and says nothing about advertising, de-identified data or data licensing. Nothing says every deployment is the same. |
Initial coding: Claude Sonnet helpers, compiled by Claude Opus 5.5 (Anthropic). Selective review and corrections: GPT-6 Astra (OpenAI). Editor: Hugo Campos. Acceptance of corrections is not product verification.
Checklist rules · Independent review and limitations · Later evidence follow-up
Earlier v1.3 assessment and control finding
Historical assessment under different rules. It does not set the current checklist zone.
Agency finding, before method v1.3
Mixed, institution-led and directly patient-facing
Legacy editorial placement, pre-method: 38
This profile has not yet been reviewed under the published method. Its finding and any number are editorial placements made before the method existed.
- Evidence
- Documentary review. Public sources only, no hands-on testing by HugoScore.
- Review status
- AI draft. No person has settled the findings.
Earlier evidence and assessment
The sections below preserve the earlier review. Its classifications and provenance are historical; the checklist above gives the current draft classification.
Earlier finding · Legacy editorial placement, pre-method: 38
Method v1.3. The control line is evidence for how the agency finding arose, not a quality score. Legacy editorial placement, pre-method: 38. Not comparable with v1.3 positions.
Mixed, institution-led and directly patient-facing
Epic Emmie combines direct patient record questions with health-system control of available actions. Sutter’s published version cannot schedule, message, or update the chart. Its chat-visibility rules do not establish an absence of retention or secondary processing.
Patient agency
How this tool changes agency
Record explanations and follow-up questions can support patients' reasoning. Feedback is useful but does not promise resolution of a disputed answer. Immediate summaries merit source and uncertainty visibility, which these pages do not establish in use.
Production models, processing, retention, deletion, automatic-summary choices and correction outcomes remain unknown. The earlier pre-release study and Reid deployment were not newly validated.
Patient agency assessment
Who sets and changes the goal?
Patients can initiate questions about their own chart and choose what to ask. The health system sets available actions and context. Sutter's explicit inability to write to the chart also means patient challenges need a separate record-correction process.
What can the patient understand, question, or do?
Record explanations and follow-up questions can support patients' reasoning. Feedback is useful but does not promise resolution of a disputed answer. Immediate summaries merit source and uncertainty visibility, which these pages do not establish in use.
Can the patient evaluate the conditions of use?
Sutter gives concrete visibility and proxy boundaries, but provider invisibility is not evidence of absent server retention or secondary use. Catalog claims must remain separate from each health system's rules.
Text findings
Conditions of use
Published controls and their limits
Sutter gives concrete visibility and proxy boundaries, but provider invisibility is not evidence of absent server retention or secondary use. Catalog claims must remain separate from each health system's rules.
What remains unknown?
Not tested or not established
Production models, processing, retention, deletion, automatic-summary choices and correction outcomes remain unknown. The earlier pre-release study and Reid deployment were not newly validated.
Who evaluated this?
AI-assisted public-source draft
Vendor statements describe published conditions, not independently verified behavior. No clinical, security, accessibility, or legal validation is claimed. Earlier evidence remains dated in the report and history.
Sources checked
- https://www.epic.com/software/emmie/
- https://www.sutterhealth.org/patient-resources/my-health-online/help-center/emmie-ai-assistant-questions
Source-specific findings and retrieval limitations are recorded in the full report.
Earlier review provenance
Criteria
CAIHL-derived HugoScore framework and September 7 qualitative review priorities. Draft v1.2 numerical anchors remain unadopted.
Reviewer
AI-assisted public-source reassessment prepared in OpenAI Codex.
AI / model
OpenAI Codex / GPT-6
Human review
Hugo Campos authorized publication of these AI-assisted draft reassessments on September 8, 2026. This does not claim comprehensive human verification of every finding.
Review date
2026-09-08
Limitations
Production models, processing, retention, deletion, automatic-summary choices and correction outcomes remain unknown. The earlier pre-release study and Reid deployment were not newly validated. No live product use, patient-data upload, account creation, code audit, clinical evaluation, or independent implementation validation.
Earlier review method
Focused public-source reassessment using CAIHL: patient authority, critical capacity, and informed control. Existing evidence plus one focused primary-source pass and at most one targeted follow-up. No live product testing. Numerical scores remain provisional editorial placements, not a new calculation.
Document-based draft · Medium for current catalog and Sutter scope (AI-assisted draft)