Large Language Model for assisting users to track costs and save lives
Product Design
Ensuring quality care in clinics and hospitals while managing costs and optimizing facility use is a challenge for busy clinical staff.
Dashboards provide data but often lack the actionable insights needed to drive decisions or implement initiatives effectively.
At the time of this product’s design, ChatGPT-4 had just become publicly available. My client sought innovative ways to deliver highly actionable insights beyond what standard dashboard interactivity or traditional business intelligence could offer.
To address this, I integrated a generative pre-trained transformer (GPT) into the workflow, which proved to be the key solution for meeting the product’s requirements.
I used example scenarios as prompts in ChatGPT-4 to generate “Drivers” within the mockup, illustrating the value of tailored insights for making critical decisions and initiating actions.
In this design, prompts would be generated by business intelligence and fed into the GPT, eliminating the need for user input and streamlining the process. Future iterations could include a prompting tool, enabling users to refine Drivers for specific cases.
Through these wireframes, I demonstrated how a Large Language Model could enhance efficiency, improve patient care in hospitals and clinics, reduce costs, and lighten the workload for clinical staff.
The “Opportunities and Drivers” screen presents the business intelligence in the “Key Indicator” section, but the real value lies in the “Opportunities” section, where ‘Drivers’ are organized within collapsible accordion panels. When collapsed, these panels reduce visual clutter, making scanning easier.
To further assist users, each panel displays a numeral on the right, indicating the number of Drivers, enabling quicker decision-making and enhancing the depth of information on the screen.
When the user selects an ‘Initiative’, a drawer panel slides in from the right, allowing the user to refine the details, assign it to specific personnel, set deadlines, and define actionable items. This creates a recorded Initiative that integrates back into the BI, enhancing the analysis of clinical operations.
The “Initiatives” screen showcases initiatives created through the “Opportunities and Drivers” interaction. Users can track progress, make edits, and generate a PDF for sharing or printed records.
Ux & Product Design, Information Architecture, Interaction Design