How S10.AI Medical Scribe improves EPIC usability

A health-system pilot showed CRUSH dramatically improved EPIC usability — cutting clicks, reducing in-basket time and giving providers their cognitive bandwidth back.

Clinician

Health-System Pilot Cohort

System-wide deployment

Practice

Regional Health System (Epic)

Multi-site, hundreds of providers

EHR integration

Epic

Live in <7 days

Compliance

HIPAA · SOC 2

BAA included

The challenge

Providers were clicking through dozens of Epic screens per visit just to chart.

In-basket volume kept growing — clinicians chipped at it after hours.

Before S10.AI

Providers were clicking through dozens of Epic screens per visit just to chart.
In-basket volume kept growing — clinicians chipped at it after hours.
Burnout was a measurable, system-wide problem.

After S10.AI

CRUSH deployed inside Hyperspace with structured write-back via FHIR.
AI Team agents handled inbox triage, refills and routine messages.
System-wide KPI dashboard tracked clicks, in-basket time and burnout signals per site.

Strategic implementation

How S10.AI was configured for this practice — without changing the clinician's workflow.

Phase 01

CRUSH inside Hyperspace

Deployed inside Epic Hyperspace with structured FHIR write-back across multi-specialty sites.

Phase 02

AI Team for in-basket

Agents triage inbox messages, draft refill responses and handle routine outreach before providers open the basket.

Phase 03

System KPI dashboard

Click counts, in-basket minutes and burnout signals tracked per provider, site and specialty.

Specialty-specific wins

What this team got that a generic AI scribe couldn't deliver.

Clicks per encounter down 46% across IM and specialty sites.
In-basket minutes per day down 38% within six weeks.
Provider NPS up 18 points pre vs post deployment.

The measurable shift

-46%

Clicks per encounter

Measured across a 6-week post-deployment window.

-38%

In-basket minutes / day

AI Team triaged and drafted responses before providers opened them.

+18%

Provider NPS uplift

Measured pre- vs post-deployment via internal survey.

What the research says

Ambient AI and EHR usability

Health-system trials of ambient AI consistently report 40–50% fewer clicks per encounter and meaningful reductions in cognitive load during visits.

Source · NEJM Catalyst

In-basket burden and burnout

In-basket time is one of the strongest predictors of physician burnout. AI triage and drafting consistently recovers 30–45 minutes per provider per day.

Source · Annals of Internal Medicine

EHR & safeguards

Integration

Epic — Hyperspace + FHIR write-back, deployed across multi-site, multi-specialty cohort.

Safeguards

SOC 2 Type II, HIPAA, HITRUST aligned. Clinician sign-off on every note and message draft.