Military Benefits Assistant uses artificial intelligence to make military benefits more accessible. I created it in the summer of 2024 with a ChatGPT account, an online Python course, and plenty of curiosity.
I had no background in technology or coding. I was a Coast Guardsman with nearly four years at sea and one year of law school under my belt. But I was signed up for Suffolk Law’s Legal Innovation and Technology Clinic in the fall and I wanted to become familiar with the subject matter. So I created a ChatGPT account, enrolled in an online Python course, and began experimenting with large language models (LLMs).
What started as curiosity quickly turned into a vision of a tool that would use artificial intelligence to make military benefits more accessible to those who have earned them. The idea that would become Military Benefits Assistant was born!
Identifying the problem
Service members, veterans, and their family members have unique opportunities to earn education degrees and professional certificates at no cost, or at a significantly discounted rate. I took advantage of Tuition Assistance to earn my master’s degree in 2022. The program paid up to $4,500 per year toward any academic degree program. But because I did not know how to access the benefit properly, I ended up paying full price for my first semester. The process was not difficult, but I had to figure it out for myself.
Then, during my tour as operations officer aboard a Coast Guard cutter in Bahrain in 2023, one of my collateral duties was serving as the education services officer. In that role I was responsible for helping crew members access programs like the Post-9/11 GI Bill, Tuition Assistance, and Credentialing Opportunities Online (COOL). It became clear to me that many service members did not know what benefits were available to them, how to determine eligibility, or how to apply.
I gave a presentation about benefits and made myself available for individual counseling meetings. I encouraged each crew member to sit down with me to discuss their education and credentialing goals. As a result, more people used a benefit, transferred one to a family member, or made a plan to do so in the future.
The lesson was simple: if service members have accessible, understandable information, they will take advantage of their benefits. My challenge was scale, now to provide similar help across the armed forces?
First steps: experimenting with AI
I saw AI as a tool that could help increase access to benefits. The complexity of military benefits is due to confusing processes, hard-to-navigate websites, lack of training, and limited awareness—especially of spouse and family entitlements. This is exactly the kind of challenge AI can help simplify.
I created a custom GPT (chatbot), trained on federal education benefits, that was able to help users figure out:
- If they are eligible for benefits
- How to learn more about the benefits they are eligible for
- How to apply for benefits
The custom GPT could converse naturally and provide information, so I bought a domain and used ChatGPT to help me build a website to host the chatbot so people could use it without logging into ChatGPT. This also helped with the trust gap among people who are skeptical of AI (or ChatGPT specifically).
I recruited some people to help stress-test the benefits tool, which exposed some significant issues:
- Sometimes it would “hallucinate” details about benefit programs or give people the wrong links
- It struggled with unexpected inputs from users
- If someone didn’t know the “right” questions to ask, the chatbot could not help them
That’s where I was when I started classes in Fall 2024.
The breakthrough: online guided interviews
My experience in the LIT Clinic gave me a new perspective.
Online guided interviews are structured, question-and-answer formats that funnel users toward relevant resources without requiring them to know the “right” question themselves. I selected the questions, and users just had to submit their information to learn about benefits they might be eligible for.
That clicked. A guided interview, not a chatbot, was the right model for Military Benefits Assistant. I assembled an AI “team” to help me build it.
I prompted Claude to generate code, then used ChatGPT to refine and optimize it for deployment on Vercel. I even created a simple prompt generator (with AI help) to strengthen my prompts.
I was able to accomplish in days and weeks what would have taken months or years. Military Benefits Assistant evolved from a fragile chatbot into a guided interview platform, a step-by-step questionnaire that matches users to likely benefits, provides links, forms, and instructions, and empowers them to take the next step. It does not replace official eligibility determinations, but it removes the first, most intimidating barrier: knowing where to start.
Results so far
Military Benefits Assistant has attracted over 1,700 visitors and over 300 site subscribers. This data shows there was a need for a tool to bridge the knowledge gap around military benefits.
Beyond the data, the true “feel good” stories are the LinkedIn messages and website feedback forms from total strangers that say things like “Thank you. This tool helped me/my spouse access our benefits.”
Networking, user feedback, and stress-testing continue to shape the development of Military Benefits Assistant, and every iteration brings me closer to my goal: increasing access to benefits for active-duty members, veterans, and families across all branches.
Businesses and nonprofits that support the military community are leaning into technology to increase access to their programs. Some have reached out to Military Benefits Assistant for help, like Coast Guard Mutual Assistance (CGMA). CGMA is the official nonprofit of the United States Coast Guard that provides financial assistance to Coast Guard families during times of need.
I built a chatbot like the original version of Military Benefits Assistant called the CGMA Program Finder that matches users with the appropriate program that CGMA offers. The tool allows users to describe their need to the bot and be directed to appropriate programs and resources.
I also created the Military Benefits Assistant Podcast to explain military benefits to the military community. The podcast has episodes “hosted” by NotebookLM AI bots as well as my conversations with people committed to increasing access to military benefits, like the President of Armed Forces Mutual, Mike Meese, U.S. Army Brig. Gen. (ret.), and the CEO of CGMA, Brooke Millard, U.S.C.G. CDR (ret.).
Thanks to AI, military members can use Military Benefits Assistant to help them access the resources they have earned through service, whether by completing a guided interview or listening to a podcast.
Lessons learned
AI is not a magic bullet. A simple prompt to ChatGPT will not yield a solution to a complex problem. A series of sophisticated prompts utilizing multiple tools paired with an innovative mindset creates an opportunity for solutions. I learned that technology is only as useful as its alignment with a user’s needs. In this case, the structure that guided interviews met users’ needs more than an AI chatbot.
Fail and iterate fast. Starting with a custom GPT, then moving to a website chatbot, then pivoting to guided interviews was not wasted effort, it was necessary trial and error. Each “failure” pushed me toward what would. With each step, I learned more about military benefits, technology, and AI.
No-code and low-code tools change the game. Without AI, I would still be stuck in a Python lesson, unable to bring my idea to life. AI tools allowed me to build, test, and launch a working site in a matter of days. This is revolutionary for non-technical founders.
Eye on the prize. My goal has always been the same: increasing access to military benefits. Technology is the vehicle, not the destination. I made every decision based on whether it would help service members and their families. Finding new ways to use technology to achieve this goal just helped me achieve it.
Looking ahead
I could not have done any of this without Suffolk Law, my experience in the LIT Clinic, and in the Generative AI and the Delivery of Legal Services class. Through them I learned to think in a whole new way and build innovative solutions with technology. I also learned skills like project management, customer experience, and design. As I continue to take classes in the LIT concentration this semester, I have no doubt that my innovative mindset will continue to expand.
Military Benefits Assistant is just getting started. The guided interview model will continue to expand, covering more programs and benefits. I plan to integrate more user feedback, refine the interface, and explore partnerships that can scale the impact across the military community. The long-term goal is to provide a platform where users can access every benefit and resource—federal, state, tribal, nonprofit, corporate, small business, etc.—through Military Benefits Assistant. This will require a large database, and most importantly, I will need to ask the right questions of users.
What began as an experiment in coding has become a mission-driven project: harnessing AI to make military benefits more accessible. If nineteen crew members on one ship could change their futures with the right information, what might be possible if we gave that access to everyone?