Enhancing Efficiency with AI-Driven Search and Summarization
We predict implementing AI-enhanced search and summarization will reduce time and effort, improve accuracy, enable faster decision-making, and drive cost savings for our Hayes customers. Ultimately enhancing symplr’s organizational efficiency and giving symplr a competitive advantage and AI pattern across product pillars.
About the problem
In the complex landscape of healthcare policy and appeals, professionals at Hayes—including policy writers, clinicians, appeal specialists, and analysts—navigate a time-consuming and error-prone process to find, extract, and summarize critical information. The existing search functionality fails to meet their needs, requiring them to manually sift through multiple documents to complete their tasks. This inefficiency not only slows productivity but also increases the risk of inaccuracies.
To address these challenges, our team explored whether AI-powered search and summarization could streamline workflows, enhance accuracy, and reduce the manual effort required for policy updates, coverage verification, appeals review, and product line comparisons. This case study examines our approach, findings, and the potential impact of AI-driven solutions in transforming the way Hayes professionals work.
My role part 1 -
Facilitate discovery + vision
In this initiative, my role was to lead the discovery phase to understand the key use cases for integrating AI into the Hayes customer workflow. I facilitated a workshop with design, research, product, engineering, marketing and customer facing teams to review Hayes users' needs, challenges, and brainstorm ideas for how AI could address those pain points. From there, I crafted a vision statement outlining the potential AI solution and customer experience, detailing how we could solve these challenges.
The final portion of the workshop was a design jam, where participants sketched ideas for AI-driven solutions to improve user workflows. After the jam, we voted on the best design based on three criteria: user delight, business viability, and engineering feasibility within a short timeline.
Discovery Workshop Board
Vision Statement
Design jam
Effort/Value Matrix on our design jam ideas
My role part 2 -
Create an execution strategy and milestones
We had a clear vision and a compelling use case to design for, but the next step was determining what we needed to design to meet that vision and achieve key milestones. It was clear that for Phase 1, we needed to deliver a quick, valuable design—something that could be built and tested with customers within just a few months. However, to stay focused, Phase 1 would need to prioritize core functionality, keeping it simple yet scalable for future phases. Each subsequent milestone would build upon Phase 1, refining and improving the user experience.
The primary goal was to get Phase 1 out the door quickly so we could test and learn: Would our users adopt this solution, and did we hit the mark with their needs?
My role part 3 -
Work with the UX Designer to bring the idea to life and show what the user experience could look like.
Milestone 1 (what we released): Roller skates
Features include: Summarization, Record feedback, Separate tab
Milestone 2: Electric Scooter
Features include: Summarization, Record feedback, Rate results, Save results + chat history
Milestone 3: Sports car
Features include: Summarization, Record feedback, Rate results, Save results + chat history, Prompt helper / question builder
Rollerskates
Electric Scooter
Sports Car
Watch it work!
Click the video above to watch me interact with our shipped AI Phase 1 solution.
Early Adopter Feedback
Early adopters are uncertain about fully trusting the AI solution at this stage. They appreciate having sources but would prefer citations that directly link to the specific section where the information was referenced. Additionally, users have expressed a preference for AI-generated answers to appear at the top of the page, similar to Google search results.