Beyond the Hype: Building AI With Purpose
An Interview with Andres Sepulveda Morales of Red Mage
Andres Sepulveda Morales was setting up a coaching session at a nonprofit when the woman across from him froze. She was the one who had asked him to be there. She’d told him she couldn’t afford to be left behind.
The cursor blinked. She didn’t type.
It took a few minutes of conversation to get to the underlying issue. She’d absorbed the same headlines everyone had: AI was taking jobs, straining power grids, remaking the world in ways nobody had agreed to. She’d watched people in her life build unhealthy fixations around chatbots, treating them like someone who actually cared. So when it came time to use the tool herself, something in her resisted.
Andres didn’t dismiss any of it. He just kept asking questions.
Somewhere in the conversation, a different picture emerged. This wasn’t a presence or a threat. It was software in the same way Excel was software, the same way her email was software, the same way the task tracker she’d used for years was software. Something to learn, not something to fear.
She looked up at him, and something in her posture loosened.
This is the work Andres has taken on: helping people understand what AI actually does, and earning their trust before asking them to use it.
A Different Kind of Tech Career
Andres entered tech nearly a decade ago as a web publisher for his university and eventually paved a path into a software developer position at a company that helped beverage distributors manage the software that tracked a case of beer from the brewery to the tax report. But between 2023 and 2025, he was laid off three times.
You’d think that would be enough to push someone out of the industry completely. Instead, it pulled him back to his purpose. “Part of me realized that I cared more about who I helped than what I was actually doing,” he says. “And so, I decided to stop being a cog and start building the machine.”
Red Mage, his value-driven AI consulting practice, was that machine. Its name came from the versatile caster in Final Fantasy who bridges Black Magic and White Magic: technical engineering on one side, human-centered design on the other. In practice, Andres might be building an AI-powered product for a founder one week and running a training session at his local library the next.
Clients range from a basketball coach to a skincare brand to a pitch competition in the Dominican Republic and everywhere in between. Andres’s core mission is to make AI work for the people the tech industry has historically left behind, through sliding-scale rates for nonprofits and underserved communities, workshops that cut through the hype, and solutions that adapt to how people already operate.
Andres is equally deliberate about who’s not on his client roster. If a company is jumping on the AI train without a clear sense of who they’re helping, he passes.
“I’m trying to answer questions that go beyond figuring out how to save money or boost productivity,” he says. “I’m always focused on what people are actually trying to accomplish…I want to see nonprofits using this to actually make a difference in their communities, things they wouldn’t have been able to do otherwise.”
Taxes and Trust
The more clients Andres worked with, the clearer it became that the barriers to AI adoption were socioeconomic, cultural, and systemic. If people don’t have the time to learn complex tools or the budget to hire a developer, what he calls the “usability tax,” they may find themselves locked out before they even begin.
That tax leaves certain communities behind. BIPOC individuals—and Latino and Hispanic communities, in particular—often encounter AI agents trained on English-dominant, Western data and perspectives. LGBTQIA+ communities encounter tools that were largely trained on data that doesn’t reflect their experiences or identities either. Older populations often haven’t been shown what the technology actually does (or that there isn’t another person on the other end of it). Creatives and nonprofits tend to approach AI with caution, asking whether adopting the technology will compromise values like sustainability or artistic integrity.
Trust is its own kind of barrier. Andrew Ng, a leading AI educator, has argued that practitioners need to be worthy of trust before communities will engage. That’s how Andres approaches his work. He runs sessions in both English and Spanish, and spends time in public libraries and nonprofit offices, sometimes starting by opening ChatGPT and writing a prompt. In many cases, it’s the first time people see the tool explained without jargon or the assumption that they already know how it works.
What He’s Building Toward
Andres wants to see organizations using AI to do things they couldn’t otherwise, reaching underserved communities and helping them make limited resources go further. He draws on his background in enterprise resource planning (ERP) systems to help small businesses navigate risks like data leaks and AI hallucinations, prioritizing transparent tools that help them retain ownership of their data.
He’s also building ArtGuard—a platform that helps artists understand how generative models use their work, and whether hosting platforms are actually protecting their art—with students at his alma mater, Northern Arizona University. ArtGuard will become a module for another solution Andres has built for artists (Vigri) to help them find in-person markets to apply to.
For Andres, all of his work comes back to the people AI was never built for, and whether they feel safe enough to engage with it. He wants younger people and people who look like him to feel represented enough to show up. He wants nobody sitting in a room somewhere, wondering whether their job is on the table.
The future he is working toward is one where AI is in people’s corner. “I want AI to help me live my life better,” he says. “Work to live, not live to work.”
And for Andres, that future starts in places like a nonprofit office: sitting across from someone who just needed to see AI work before she could believe in it.