Prior Authorization Appeal Tool
Powered by AI
Leveraging AI design tools in combination with usability testing to build a tool that helps patients understand and fight their health insurance denials.
Role
Product Designer, UX Researcher
Timeline
4 weeks
PLATFORM
Web
UI/UX Design
AI Design
Prototyping
UX Research
Healthcare UX
Prior Authorization Appeal Tool
Powered by AI
Leveraging AI design tools in combination with usability testing to build a tool that helps patients understand and fight their health insurance denials.
Role
Product Designer, UX Researcher
Timeline
4 weeks
PLATFORM
Web
UI/UX Design
AI Design
Prototyping
UX Research
Healthcare UX
Prior Authorization Appeal Tool
Powered by AI
Leveraging AI design tools in combination with usability testing to build a tool that helps patients understand and fight their health insurance denials.
Role
Product Designer, UX Researcher
Timeline
4 weeks
UI/UX Design
AI Design
Prototyping
UX Research
Healthcare UX
SecondLook Demo Reel
SecondLook Demo Reel
Overview
Overview
My interests in the healthcare space led me to Eric Shumake's Healthcare UX Certificate course. It was here that I had the opportunity to design, prototype and launch a tool that helps patients understand and confidently appeal prior authorization denials from their insurance. I used Claude and Anything as part of my AI-powered workflow to create and quickly iterate on a thorough interactive prototype.
My interests in the healthcare space led me to Eric Shumake's Healthcare UX Certificate course. It was here that I had the opportunity to design, prototype and launch a tool that helps patients understand and confidently appeal prior authorization denials from their insurance. I used Claude and Anything as part of my AI-powered workflow to create and quickly iterate on a thorough interactive prototype.
The Problem
Waiting on prior authorization for a medical procedure can already be a nerve-wracking experience, but if that PA request is denied by insurance, patients are often left stressed and confused.
Scientists and researchers are losing out on accurate insights due to difficulties with GeoMx spatial analysis softwares.
Real Clinical Consequences
Without insurance approval, necessary care could be delayed, leading to permanent damage.
Weak Denial Reasoning
Denials are often not clinically grounded, relying on proprietary guidelines that are inconsistently applied.
Patients Assume That's It
When a denial letter arrives in the mail, many patients don't realize that 80% of appeals succeed.
Unnecessary Stress
Patients that are denied procedures are then caught up in being sick and fighting their insurer at the same time.
The Problem
Waiting on prior authorization for a medical procedure can already be a nerve-wracking experience, but if that PA request is denied by insurance, patients are often left stressed and confused.
Scientists and researchers are losing out on accurate insights due to difficulties with GeoMx spatial analysis softwares.
Real Clinical Consequences
Without insurance approval, necessary care could be delayed, leading to permanent damage.
Weak Denial Reasoning
Denials are often not clinically grounded, relying on proprietary guidelines that are inconsistently applied.
Patients Assume That's It
When a denial letter arrives in the mail, many patients don't realize that 80% of appeals succeed.
Unnecessary Stress
Patients that are denied procedures are then caught up in being sick and fighting their insurer at the same time.
The Problem
Waiting on prior authorization for a medical procedure can already be a nerve-wracking experience, but if that PA request is denied by insurance, patients are often left stressed and confused.
Scientists and researchers are losing out on accurate insights due to difficulties with GeoMx spatial analysis softwares.
Real Clinical Consequences
Without insurance approval, necessary care could be delayed, leading to permanent damage.
Weak Denial Reasoning
Denials are often not clinically grounded, relying on proprietary guidelines that are inconsistently applied.
Patients Assume That's It
When a denial letter arrives in the mail, many patients don't realize that 80% of appeals succeed.
Unnecessary Stress
Patients that are denied procedures are then caught up in being sick and fighting their insurer at the same time.
Research & Discovery
First round of research consisted of gathering information from health forums and articles on the state of prior authorization in the United States. Usability tests were also conducted on a prototype.
2
Usabilty Tests + Interviews
6
Feedback responses
7
Secondary Sources
Key Insights
Patients are largely unfamiliar with the PA process unless they are right in the middle of it.
Most participants were somewhat aware that they can appeal a denial but cannot explain concrete next steps.
As much as 40–90% of appeals are successful, yet less than 1% of denials are appealed.
Research & Discovery
First round of research consisted of gathering information from health forums and articles on the state of prior authorization in the United States. Usability tests were also conducted on a prototype.
2
Usabilty Tests + Interviews
6
Feedback responses
7
Secondary Sources
Key Insights
Patients are largely unfamiliar with the PA process unless they are right in the middle of it.
Most participants were somewhat aware that they can appeal a denial but cannot explain concrete next steps.
As much as 40–90% of appeals are successful, yet less than 1% of denials are appealed.
Section Title
SecondLook assists the patient in understanding their denial letter, parsing through and translating the insurance jargon, and provides a framework of actionable next steps.
Denial Analysis
Patients can upload a PDF of their letter for the AI to analyze, presenting the key reasons for denial and actionable next steps.
Appeal Letter Generation
Based off the details in the letter and clinical rationale from the clinician's notes, the user can generate an appeal letter.
Step-By-Step Guide
Provides a framework a patient can follow as they move to fight back their denial. Shows next steps that include possible escalation paths.
AI-Powered
Implements Gemini API and Claude integration for PDF text extraction, content analysis, and letter generation. Prototyped with Anything.com.
Section Title
SecondLook assists the patient in understanding their denial letter, parsing through and translating the insurance jargon, and provides a framework of actionable next steps.
Denial Analysis
Patients can upload a PDF of their letter for the AI to analyze, presenting the key reasons for denial and actionable next steps.
Appeal Letter Generation
Based off the details in the letter and clinical rationale from the clinician's notes, the user can generate an appeal letter.
Step-By-Step Guide
Provides a framework a patient can follow as they move to fight back their denial. Shows next steps that include possible escalation paths.
AI-Powered
Implements Gemini API and Claude integration for PDF text extraction, content analysis, and letter generation. Prototyped with Anything.com.
Section Title
SecondLook assists the patient in understanding their denial letter, parsing through and translating the insurance jargon, and provides a framework of actionable next steps.
Denial Analysis
Patients can upload a PDF of their letter for the AI to analyze, presenting the key reasons for denial and actionable next steps.
Appeal Letter Generation
Based off the details in the letter and clinical rationale from the clinician's notes, the user can generate an appeal letter.
Step-By-Step Guide
Provides a framework a patient can follow as they move to fight back their denial. Shows next steps that include possible escalation paths.
AI-Powered
Implements Gemini API and Claude integration for PDF text extraction, content analysis, and letter generation. Prototyped with Anything.com.
Process: Initial Prompts
Process: Initial Prompts
My first prompts were experimental, lacking specificity and context. I was curious what the AI would generate given these instructions, though was not surprised that the results looked like a generic landing page.
My first prompts were experimental, lacking specificity and context. I was curious what the AI would generate given these instructions, though was not surprised that the results looked like a generic landing page.
"A consumer-side denial appeal assistant. Patient-facing appeal tools for denied prior authorization requests are almost entirely absent. 82% of appeals succeed; almost no one files them."
"A consumer-side denial appeal assistant. Patient-facing appeal tools for denied prior authorization requests are almost entirely absent. 82% of appeals succeed; almost no one files them."
Besides looking clean, this landing page is pretty useless. The content itself vaguely references some features that remain unclear in how they work and the call to action button doesn't link to anything. Even the site's visual design lacks focus, confused between being a sleek tech startup and an established healthcare organization.
On the next try, I defined three core features that I wanted to see in the information architecture. I also outlined a specific visual direction that I wanted for the tool and some technical specs.
Besides looking clean, this landing page is pretty useless. The content itself vaguely references some features that remain unclear in how they work and the call to action button doesn't link to anything. Even the site's visual design lacks focus, confused between being a sleek tech startup and an established healthcare organization.
On the next try, I defined three core features that I wanted to see in the information architecture. I also outlined a specific visual direction and some technical specs.
Besides looking clean, this landing page is pretty useless. The content itself vaguely references some features that remain unclear in how they work and the call to action button doesn't link to anything. Even the site's visual design lacks focus, confused between being a sleek tech startup and an established healthcare organization.
On the next try, I defined three core features that I wanted to see in the information architecture. I also outlined a specific visual direction that I wanted for the tool and some technical specs.
"Design direction: Trustworthy but empowering — not clinical/cold. Think of a legal aid organization that also has great UX. Use a warm, grounded color palette (https://coolors.co/8e3b46-fef8ec-fdf0d5-003049-669bbc). Typography should feel serious but human — a serif display font paired with a clean sans-serif body. The interface should feel like it's on the patient's side."
"Design direction: Trustworthy but empowering — not clinical/cold. Think of a legal aid organization that also has great UX. Use a warm, grounded color palette (https://coolors.co/8e3b46-fef8ec-fdf0d5-003049-669bbc). Typography should feel serious but human — a serif display font paired with a clean sans-serif body. The interface should feel like it's on the patient's side."
"Denial Letter Analyzer — User uploads or pastes their denial letter. The tool extracts key fields (denial reason, procedure code, insurer name, deadline to appeal) and translates bureaucratic language into plain English. Show a "decoded" summary card with a severity/urgency indicator and the specific grounds cited for denial."
"Denial Letter Analyzer — User uploads or pastes their denial letter. The tool extracts key fields (denial reason, procedure code, insurer name, deadline to appeal) and translates bureaucratic language into plain English. Show a "decoded" summary card with a severity/urgency indicator and the specific grounds cited for denial."
"Appeal Letter Generator — Based on the analyzed denial, generate a tailored appeal letter. The interface should show editable fields (patient name, provider, clinical rationale), a live letter preview, and one-click copy/download. Include a confidence signal like, This denial reason has an 87% overturn rate."
"Appeal Letter Generator — Based on the analyzed denial, generate a tailored appeal letter. The interface should show editable fields (patient name, provider, clinical rationale), a live letter preview, and one-click copy/download. Include a confidence signal like, This denial reason has an 87% overturn rate."
"Step-by-Step Appeal Submission Guide — A guided checklist/stepper that walks the user through: identifying the right submission channel (fax, portal, certified mail), gathering supporting documents (EOB, medical records, doctor's letter of medical necessity), tracking deadlines, and escalation paths (internal appeal → external review → state insurance commissioner)."
"Step-by-Step Appeal Submission Guide — A guided checklist/stepper that walks the user through: identifying the right submission channel (fax, portal, certified mail), gathering supporting documents (EOB, medical records, doctor's letter of medical necessity), tracking deadlines, and escalation paths (internal appeal → external review → state insurance commissioner)."
Looking for design help?
Let's talk
Book A chat
Currently booking engagements through May
Looking for design help?
Let's talk
Book A chat
Currently booking engagements through May
Looking for design help?
Let's talk
Book A chat
Currently booking engagements through May