What do homegrown Cincinnati companies like Graeter’s, Rhinegeist, Cintas, TQL, Gorilla Glue, Messer Construction and nonprofits like the YMCA of Greater Cincinnati have in common? While ice cream scoops, logistics, beer and kid camps are all wildly different, these goods and services are all offered by mid-size corporations, stuck in the middle and facing all the problems that come with the issues served up at scale.
Unlike small businesses that can thrive without hierarchy, larger companies require distinct departments and more formalized structure, including functions like human resources (HR), operations, marketing, customer service, sales and development (aka fundraising) for nonprofits. And different from massive global corporations with impressive budgets, a mid-size business may need to keep headcount and expenditures on systems and tools to a minimum.
The future is now
The burgeoning bumper crop of budget-friendly AI tools offer solutions that may seem the stuff of science fiction, but are readily available (and very real) right now. On the forefront of this innovation stands Jorge Perez, the president and CEO of the
YMCA of Greater Cincinnati. He’s an early adopter of an exciting AI capability dubbed “customGPTS” within ChatGPT.
For Perez, it all started with an agent named Yasmin. “The very first AI [assistant] we created was the Youth Development Advisor. I sat there and created a youth development expert with all the values of the YMCA and all the things that I wanted every youth development leader to have and to know and all the adaptability for how you work with kids that are under-resourced versus kids in suburbia. And all the DEI stuff was baked into that AI. And then we started asking it questions and it was shocking,” recalls Perez.
![]()
Jorge Perez, president and CEO of the YMCA of Greater CincinnatiUsing his $20/month paid subscription to ChatGPT, Perez was able to create and train Yasmin himself in less than one hour, using all the existing materials he had created during his time leading YMCAs in communities across the country.
AI assistants offer 24/7 accessibility
You can give Yasmin a prompt like, “‘Help me create a STEM-based program for seventh grade girls,’ and it was wonderful,” says Perez. “Or help me understand how to improve the impact of this summer camp program for the kids. And it could walk you through what you needed to do,” he explains. Instead of navigating a website or a complex file sharing system to hunt down information, youth development coordinators can now just ask Yasmin. “Yasmin gives you the answers, and it doesn’t give generic answers, it gives you curated answers that we taught Yasmin,” he says. And that was really just the beginning.
Now there are 16 total AI assistants linked to the most important functions at the Y.
- Bailey: Board Development
- Carter: Crisis Communication
- Casey: Contract
- Dakota: Diversity, Equity, Inclusion + Belonging
- Finley: Fundraising
- Flynn: Facilities
- Frankie: Finance
- Harmony: Human Resources
- Izzy: IT
- Marlow: Membership
- Morgan: Marketing
- Skyler: Safety and Risk
- Sterling: Strategy
- Valor: Volunteer
- Yanis: Youth Sports
- Yasmin: Youth Development
Each of these AI assistants embodies decades of information and institutional knowledge, always at the ready to not just search and retrieve answers (like search functionalities), but to use that source material to create actual usable plans tailor-made to fit the user’s exact needs. They’re an answer to several painful problems that have long plagued YMCAs in specific, and many nonprofits in general. Due to high turnover and limited hiring budgets, keeping employees up-to-date on training has always been a fairly impossible task. Until AI.
The Y’s training problem
![]()
Aaron Bozsan YMCA of Greater Cincinnati, Campbell County localionAt the national level, the YMCA has as many as 150,000 part-time workers, with about 40% turnover every year. Plus, full-time employees. “It was impossible for us at that office to keep training all these youth development leaders. I couldn't produce enough training to get in front of those individuals. And so, we were trying to create videos and stories and trying to see if it can motivate people to understand the basics of youth development. And then AI gets created. So that's why we did it. We said, ‘Well, let's try this.’ And then we created one platform and then we tried another one and another one,” he says.
Each of these assistants can interact with 100,000 different users simultaneously, effectively multiplying training capabilities almost infinitely. Of course, that’s if you can successfully get staff to adopt the technology.
The big leap is switching seats
The main barrier? Utilizing this incredible technology requires a leap, a mindset shift Perez thinks of as “switching seats.” Many of us have spent our lives as doers, or what Perez calls “creators.” And now we need to become what he describes as “curators.” Sliding into the role of “curator” can present problems, especially for those who have a limited understanding of the underlying capabilities. For example, if you don’t understand the outputs produced by programs within Microsoft Suite, you cannot direct AI to harness those powers to the fullest extent. Also, it requires an understanding that AI assistants are different from a Google search—rather than answering questions, they’re providing curated outputs.
Now (re)introducing your AI assistants
At the Y, Perez is embarking upon a reintroduction campaign for his 16 AI assistants.
“We've created very short videos. So sometimes we do short videos on how to use. Like right now I'm sending an email a month, and in February, we reintroduced Bailey, the board advisor and said, ‘Hey, here's a short one-minute video on what Bailey can do, and please ask it these questions right now.’ Because when I released 16, it was too much. Their brains exploded,” he recalls. This simplified approach triggers in-the-moment, hands-on learning where users get to experience the magic that happens when they become capable curators.
“The way the agents work is they eliminate the need to learn how to write a prompt,” explains Perez. Since each assistant is already deeply trained in a specific area, there’s no need to “set the scene” and provide the type of background required for the general-usage area of ChatGPT, Claude or Microsoft Co-Pilot.
Here’s what comes next…
If you ask Perez, this is only the tip of the iceberg. What comes next? It’s called agentic AI, and it represents another leap forward in the trajectory of AI becoming an embedded part of the workforce of the future. Agentic AI assistants can take initiative and act independently of their curator.
“The next version of Yasmin will be self-automated, and will go around the country and find all the youth development people and say, ‘Hi, I'm Yasmin, how can I help you?’ And then check in with people on Mondays or find them posting things on social media and say, ‘Hey Jessica, I saw you're launching a program for these kids. How can I help you?’ And I think that's what comes next already.”