AI innovation for Healthcare Policy & PR
Concept-to-MVP: reimagining healthcare workflows with generative AI
Note: to ensure client confidentiality, some details have been redacted. Please get in touch if you would like to learn more!
Overview
Under a provincial directive to modernize internal operations, a public healthcare network partnered with IBM. I collaborated with the IBM Client Engineering team to incubate two generative AI use cases and develop functional MVPs that demonstrated the potential for broader implementation.
Official portal and news release:
1) Identify key operational issues in healthcare services
2) Prioritize two GenAI modernization use cases
3) Deliver 2 MVP solutions within a 1-month timeframe
I led workshop activities and facilitated conversations with healthcare workers. This helped inform the experience strategy and scope requirements for development.
Nour Al-Safi, Leandra Keller, Stephanie Duguay, Maya Michniewicz
About the client
Our client is a provincial health authority in Canada. They are the largest employer in the province with more than 23,000 employees and 2,500 physicians working from 45 different facilities.
Photo by Piron Guillaume on Unsplash
Prioritized challenges
Conversations with healthcare professionals highlighted key issues that resulted in 6 different use cases. We prioritized the top two:
Nurses needed a way to quickly search and understand existing policies. There were 4,000 internal policies across different webpages and documents — some of which were redundant or outdated.
A Retrieval Augmented Generation (RAG) implementation for front-line healthcare staff to quickly lookup and verify policy and health care directives
Delayed & strenuous communications
At the CEO's office, the communications team faced challenges developing content from scratch. Existing templates lacked the necessary nuance and tight deadlines further complicated the process, leaving little time for final reviews.
Generation and conversion of communications materials of various types (e.g. key points, news release, speaking notes, etc.) from different viewpoints (Premier, CEO) to various audiences (clinical vs. general public)
Policy search tool
I worked with AI developers to scope out the most high-impact features for each of the use cases. We developed and showcased two MVPs. Here is a short snippet of one of the demoes:
Policy maintenance
The Gen AI solution becomes a central repository for all users, substituting three current legacy systems. Front-line healthcare staff can quickly lookup and verify policy and care directives, while managers can continuously maintain and update evolving policies.
from idea to solution
Through focused efforts, we crystallized the top use cases. Central to our approach was cross-functional collaboration to ground our work in the real needs of front-line hospital employees.
key activities
Client briefing to explore applications of Generative AI in healthcare
Use case exploration and ideation session with stakeholders
Use case prioritization workshop with business and technical teams
Interviews with would-be users (e.g. nurses) and proto-persona development
IBM Design Thinking sessions for solution scoping and feature prioritization
AI experience design and prompt engineering
Co-designing with real users
Drawing from persona pain points, we led mind mapping exercises—in this example, with nurses and policy managers—to identify opportunity areas and prioritize MVP vs. long-term features.
Where prompt engineering meets design
We ensured that AI responses were both context-aware and user-centric. We translated user intent into structured prompts that integrated system-level instructions, user query context, and business logic to guide the AI’s response.
continued partnership
Our client was eager to scale the two MVP solutions, leading to ongoing collaboration and more innovation-based endeavours on the horizon.
Save 100,000 hours per year, leading to 31% time reduction spent on policy-related tasks
Save 5.9 hours per communications request, leading to 83% improved response time
Tools were recognized as a significant improvement over the existing tools by our client
Our client decided to scale the two MVP solution across the hospital network to other teams