Gov AI
Serving underserved populations with conversational design that help to improve access to state benefits
January 2020 - Present
UX Researcher / Designer
Participatory Workshop / Conversation Design / Contextual Inquiry / Observation / Guerilla Research
Overview
New Update Keep up with us on Medium! Check it out here 👩🏻💻🌟
Client Gov AI
Team Simran Jobanputra / Laura Fulton / Tommy Byler / Conlon Novak
Story
Gov AI is a start-up that aims to create applications of AI technologies to government services including Supplemental Nutrition Assistance Program (SNAP) and state benefits access tool to increase accessibility of these programs for citizens who are applying for government assistance while simultaneously delivering value to the policymakers.
As the lead researcher, my role is to lead foundational and ethnographic research to understand the sociological and economic barriers that hinder people from easily accessing government services and the needs of underserved populations within the existing state benefits ecosystem.
To design a CUI that simplifies the state benefits application process, we are conducting analogous domain research and generative research in parallel, including on-site contextual inquiries, observations, and guerrilla research with stakeholders to understand the varying factors we must consider when situating a conversational agent within the government information technology infrastructure and to fuel ideation for a multi-touchpoint integration.
PROBLEM
Current status of state benefits ecosystem
Reading through articles and client’s previous documentation of secondary sources gave us a better understanding of the current state of state benefits ecosystem. As a way for our team to build empathy with those who are left behind when applying for benefits, we conducted informal interviews with those who have personal experiences of applying for assistance programs including medical expenses and homeowner’s insurance.
The Atlantic and online COMPASS tool for access to benefits
We gained preliminary insights into what the process is like for someone taking over applications for their loved ones who now struggle with applying for benefits. With this, I framed our research direction to focus on foundational and ethnographic research to dive deeper into the unmet needs, attitudes, and driving motivations of people who are in need of assistance and discover new design opportunities by uncovering different points of entries when people apply for state benefits.
Creating our own internal repository of secondary sources (on right) and learning through client’s documentation of secondary sources (on left)
TIMELINE
Working in sprints
The current timeline for this project is running on sprints of two-week cadence so that we can support design and development efforts. During the first few weeks, I performed desktop research to learn more about the existing state benefits ecosystem and planned kickoff meeting with the team to demonstrate how research shapes our design direction. I devised a plan to have parallel research processes, splitting into initiatives focused on on-site contextual inquiries and observational studies with guerilla research to get a better picture of our target population.
WORKSHOP
Kickoff meeting
For the kickoff meeting with the client, I wanted to introduce the kickoff as a co-creating workshop and plan out activities that introduce the clients to the design thinking process.
Our team planned out activities that aim to challenge assumptions and identify knowns and unknowns, including:
Six thinking hats
“Putting on” a series of different ways of thinking in the form of real, colorful hats to help approach a topic from every angle.
We formed clusters based on what information is missing, what we want to know more about, how we feel about the idea, feasibility of ideas, and other risks associated with different use case scenarios.
CUI mockup demo
Using VoiceFlow to mockup a CUI demo as a topic to ideate about for the Six Thinking Hats activity to identify gaps in situating conversational agents alone in state benefits ecosystem and discover new avenues to explore. This helped us to enable clients to not hone in on a particular solution yet.
Reversing assumptions
Intentionally flipping preconceptions we have about the problem space in order to brainstorm novel ideas and opportunities by walking them through existing assumptions, reversed assumptions, and ideating new service ideas from the reverse.
Some parts are blurred for confidentiality
Stakeholder mapping
Building out the relationships and value flow exchanges between varying groups of stakeholders to explore who is involved in the state benefits ecosystem. This helped us identify key players, including federal agencies, state programs, and private service providers that we had on our radar heading into user research. We hung this up in our office to have a shared understanding as a team!
Major stakeholder groups that emerged from this activity include:
- Non-profit organizations within Allegheny County
- Government entities including federal agencies, state programs, private service providers
- GovTech startups
- For-profit organizations including venture capital firms
- Citizens including unemployed, homeless, disabled people, single parents,and seniors
Some parts are blurred for confidentiality
ANALOGOUS DOMAIN RESEARCH
Shifting focus to a new context
Our team recognized that connecting to underserved populations through community-based organizations, non-profits, and exploring analogous domains is critical for us to empathize with users, identify unmet needs, and uncover key insights.
We brainstormed words that describe the application process for benefits:
- Time-sensitive
- Immediate needs
- High Complexity
- Lack of assurance
- Lack of assurance (no guaranteed outcome)
- Potential mismatch in screening vs. outcome
- Lack of transparency into why the decision is made
- Decision made by someone else
- Binary decisions
- Limited resources
This led us to formulate a driving statement which guided us to identify analogous domains:
To shift our focus to a new context and explore the problem space through a different lens, we identified these domains:
ON-SITE CONTEXTUAL INQUIRIES
Using contextual inquiry to learn about non-profit organizations
To better understand the relationships that other stakeholder groups share with people in need, I am conducting on-site contextual inquiries with non-profit organizations to learn about their experiences and limitations they face when helping people who may need benefits through a mix of interviewing and observations.
Snippet of guide (on left) and spreadsheet of contacts for non-profits who have filled out our interest survey (on right)
We have reached out to many non-profit organizations who help people who may need state benefits and are situated between the state government and people in need. These organizations range from engaging in community outreach for people in crisis to giving access to quality, affordable health care and food. Being able to visit them on-site and talk with them while watching their day to day processes would help us understand the needs of our stakeholders better. I am also conducting ride-along studies with truck drivers who are delivering food to various non-profit partners so that we can better understand the handoff process in the field.
Some of the non-profits that we have scheduled for so far include:
- Greater Pittsburgh Community Food Bank
- Consumer Health Coaliton
- 412 Food Rescue
Visiting on-site at 412 Food Rescue during planning phase to build rapport with people
OBSERVATIONS
Planning observational studies
We are planning to immerse ourselves in areas that are situated between state governments and people in need, to observe people withiin natural settings.
Some areas that we have planned for are:
- Pittsburgh Social Security Administration office
- Pennsylvania Department of Human Services
GUERRILLA RESEARCH
Running quick guerrilla research in parallel
Because we expect contact with non-profit organizations to take time, we wanted to perform quick, guerrilla research to run in parallel so that we can build empathy for the populations we are trying to serve by outreaching for people first-hand.
Areas where we have planned to run quick research:
- Giant Eagle (high foot-traffic grocery store that supports SNAP)
- Target
Our team (Conlon on left and Laura on right) visiting Giant Eagle to get a sense of the space during planning phase
OUTREARCH
Connecting with GovTech start-ups and government entities
We are continuously trying to build rapports and lasting relationships with our primary stakeholder groups and end-users so that we can co-create meaningful value within the state benefits ecosystem. I’m also creating flyers attached with screeners so that we can outreach to people who have had experience with or have interest in applying for benefits through community bulletin boards.
