Stories shape everything about how we experience the world. They’re the way we live, share, and learn. In an era where artificial intelligence influences more and more aspects of our daily lives, stories remind us of something crucial: technology may be sophisticated, but at the end of the day, it exists to serve people.
That’s why I’m starting this series—a space to explore the human side of AI-powered experiences through the lens of UX research. For all the innovation and complexity behind AI systems, real people with real needs often find themselves trapped by the very tools designed to help them.
This is one of those stories.
When the Algorithm Decided Adam Was a Spammer
Meet Adam (not his real name), a senior hair stylist and a close friend who recently took the leap to open his own business. Like many entrepreneurs, things moved fast—faster than expected. He had his domain name secured and was racing to set everything else in motion. Suddenly, he had just a few hours to let his clients know about his big move.
Not wanting to use his personal number for business, Adam quickly got a VoIP line and began texting his client list. What he didn’t know? Sending a high volume of texts in a short time is exactly the kind of behavior that spam-detection algorithms are trained to block.
Within hours, Adam’s number was flagged. His messages never reached his clients. Panic set in.
The Bigger Picture
Adam’s case isn’t unique. Across industries, small businesses are running into the invisible walls of algorithmic decision-making.
AI systems designed to protect us from fraud or abuse are inadvertently harming legitimate users. Worse, they often do so without transparency. Thresholds, data sources, and carrier rules intersect in ways that most entrepreneurs don’t even realize until it’s too late.
What’s left is a confusing maze of restrictions that ordinary people—often at the most critical moments of their professional journey—have no idea they’re walking into.
A Human-Centered Solution
So how do we, as UX researchers and product teams, build systems that serve people first?
Here’s a four-step framework I use to keep humans in the loop:
- Design well
The user interface should be the first line of defense. Clear warnings, visual indicators, and plain-language explanations can change outcomes. Imagine if Adam had seen: “Sending 100+ texts quickly may trigger spam filters.” He would have paced his outreach differently. - Evaluate everything
Test concepts with real users—not only for usability but also for trust and comprehension. And don’t stop at interfaces. Machine learning models require ongoing evaluation, especially for edge cases like Adam’s that are rarely represented in training data. - Context is king
Early research matters. We must understand the full spectrum of user scenarios before we build. Mapping urgent, high-stakes transitions (like Adam’s sudden business move) against real personas and mental models could have surfaced this risk earlier. - Reboot org culture
At the heart of this challenge is organizational culture. How do companies balance speed with responsibility? What guardrails ensure that research insights don’t get lost between teams? And how do we keep learning from real user stories rather than treating them as anomalies?
The Bottom Line
Every algorithm touches real people with real dreams. Behind every data point is a story like Adam’s—a story of ambition, risk, and resilience.
As creators, researchers, and innovators, our responsibility is clear: never lose sight of the human on the other side of our decisions.
What about you—have you experienced your own “Adam moment”? Share your story. Together, we can design systems that not only perform well but also respect and empower the people who rely on them.
Interested in more? Check out Renu Zunjarwad post: https://www.linkedin.com/pulse/humans-loop-rz-renu-zunjarwad-ph-d–z8ksc/
More of our blog entries in: https://blog.cedim.edu.mx/