Inspiration : Navigating EHRs in real time is complex and time consuming leaving physicians with less patient time and revenue, so they are prevented from recording, retrieving and analyzing patient data at the point of care necessary to make smarter clinical decisions and get better outcomes.

What it does: Mili, the Artificial Intelligence powered virtual clinical assistant understands a physician's naturally spoken commands to instantly retrieve and analyze patient data to provide key data driven insights for accurate diagnosis and choose the right treatment plans. This generates better outcomes with smarter decisions, lower costs and fewer errors.

How we built it: A cloud & SaaS platform based middleware that connects to an EHR to access patient data and provide instant analysis to Physicians based on their voice commands delivered through their iPhone (iOS) mobile app.

Challenges we ran into: Getting Mili to reliably understand a physician's voice commands without the physician having to remember any commands, command syntax or keywords.

Providing the required user experience in terms of uptime and response time for speech service especially when user limited to 3G wireless service

Accomplishments that we're proud of: Over 23 Healthcare provider organizations have adopted our application for use by their clinicians and are saving over $3000/Mo/Clinician by eliminating Scribe service. Their patients are spending an average of 18% less time in Skilled Nursing Home compared to national average which equates to approximately $2500 in savings per episode of care.

What we learned: Getting an application production ready takes lot more effort than just making it bug free.

What's next for Mili -The Virtual Clinical assistant: The ability to diagnose and identify the right treatment plans based on the clinical guidelines. Eventually Mili will be able to anticipate a patient's deteriorating condition by mining vast clinical data to uncover hidden trends and connections that humans miss to suggest a preventive measure instead of a treatment plan.

Built With

Share this project: