

Social Benefits
Designing an AI-enhanced social benefits application to streamline requests and fulfillment
Products
ServiceNow Platform
CSM Workspace
Public Sector Digital Service
Playbooks
TEAM
Ryan Griffin - Product Owner
PSDS Engineering Team
CSM Playbook Design Team
CSM Mobile Design Team
Duration
Two weeks
(June 2023)
My Contributions
-
Led two week design sprint and strategy for the Social Benefit concept work
-
Designed an end-to-end workflow that integrated AI-enhancements
-
Delivered and presented a working prototype and story for leadership
Outcomes
-
~40% faster constituent application process via AI-assisted intake and document scraping
-
Improved agent decision-making through case and policy summarization
-
Secured leadership buy-in for AI integrations for future workflows
Opportunity
Problem
Government agencies face challenges delivering consistent and seamless experience for applying to social benefits. This application process is often complex and lengthy with frequent request for redundant information and a lack of clear guidance or support for both constituents and agents.
Key challenges
-
Our enhancements needed to compliment work from other teams within our business unit
-
Complex and frustrating application process causes frustration for constituents
-
Agents spend considerable time looking for fraud during benefit intake
-
Large agent workloads with no summarization lead to slow decision-making and increased risk of errors
Discovery
Applied research
Due to the short time frame for this project, I worked with the product manager to identify multiple social benefit experiences; deciding to focus our story on a child care payment.

Tennessee child benefit application site
Persona development
To ensure our narrative accurately reflected the social benefits experience, I developed detailed constituent and agent personas that highlighted key pain points from the application processes.

Constituent and case agent personas
Journey mapping
I identified 3+ high-friction points for applicants (confusing intake, difficult verification process, and redundant data requests) and 2 key agent pain points (document review and policy summarization).

Final updated journey
Possible solutions
Based on my preliminary research I identified a 3 opportunities that were not being utilized by other industry teams:
-
Streamlined constituent intake forms by leveraging AI-assisted data extraction from submitted documents
-
Integrated verification system to minimize agents’ manual efforts in fraud detection
-
Implemented AI-powered summarization tools to enable agents to efficiently review and assess cases
Design
AI-driven intake
I designed an AI chatbot to create a frictionless, approachable intake experience that guided applicants through a series of questions that fulfilled required information and identified essential documents.



Virtual agent extraction
Identity verification
I streamlined identity verification by embedding ID.me authentication, enabling secure data submission, automated checks against trusted databases, and reducing manual reviews while minimizing fraud risks.

ID.me log in screens
Document scraping
I streamlined the application process by integrating intelligent document scraping and conflict detection, allowing constituents to validate and resolve missing or conflicting data in real time.



Document extraction
AI summarization
I enhanced the agent experience with AI-driven case and policy summarization, delivering instant, non-disruptive context to support rapid decision-making during frequent context switches.

Case and policy summarization
Presentation
Leadership Presentation
I collaborated with my team to refine and deliver a demo of the experience for our leadership which was extremely well-received. This presentation shifted our leadership's mindset around how governments could leverage AI-capabilities and helped drive more funding to the development of AI-capabilities in the playbook framework.
Presentation flow for leadership
Outcomes
In two weeks, my team and I were able to deliver a social benefit experience that showcased:
-
~40% faster constituent application process via AI-assisted intake and document scraping
-
Improved agent decision-making through case and policy summarization
-
Secured leadership buy-in for AI integrations for future playbook workflows


My Learnings
As lead designer, I gained confidence guiding the team from concept mapping through to executive presentations. I strengthened my ability to communicate effectively by aligning stakeholders around a clear user journey that unified design and storytelling. I also learned how to demonstrate AI’s tangible benefits in ways that directly addressed user pain points, which improved cross-functional alignment.
I found it especially rewarding collaborating with other teams to explore how unique AI capabilities could deliver meaningful value. While not all of these concepts were showcased in the final presentation, these conversations created visibility for these teams about playbook capabilities and opened opportunities for future adoption.