Designing AI-Enhanced UX Systems
This summer, I joined Roamly as a contract UX designer to help redesign their marketing website alongside a talented visual designer. While our work focused on improving the user experience and conversion across key pages, something bigger was happening around us: Roamly was undergoing a full AI transformation.
The company was actively exploring how to embed AI into internal operations and workflows. Not just customer-facing tools, but internal systems too. As I listened in on strategy sessions and hallway conversations, I found myself thinking: What would it look like to bring UX strategy to that layer of the business?
From UX to Systems Strategy
I started using Gemini to support and scale parts of my UX workflow. While I’m still in the early stages of exploring its full capabilities, it’s already become a helpful tool for thinking at a systems level.
Here’s how I’ve been using it so far:
Auditing live website flows to surface inconsistencies in tone, structure, or hierarchy, especially across high conversion pages
Synthesizing stakeholder input, user research, and business goals into clear and actionable design priorities
Generating first draft UX copy and SEO aware content frameworks to guide collaboration with the copywriter
Exploring how custom Gems might support internal systems, such as generating UX documentation, auditing design libraries, or assisting with content tagging
The more I work with Gemini, the more it feels like a collaborator in systems thinking. It’s not replacing the design work. It’s strengthening the connective tissue between people, tools, and processes.
Designing with Trust and Usefulness in Mind
One theme that kept coming up: trust. If companies are going to bring AI into their operations, the systems can’t just be efficient. They have to be understandable, predictable, and easy to override.
I began thinking about how UX strategy could support this:
Building prompts and checkpoints that keep humans in the loop
Creating workflows that are augmented, not automated blindly
Designing signals that help teams know what AI is doing (and why)
These weren’t product-facing UX decisions. They were system-level ones. And they felt just as important.
A New Direction for New Friend
This experience at Roamly sparked the beginnings of a new service offering I’m exploring for New Friend: UX Systems Strategy for AI Integration. It’s still in development, but the core idea is simple: Help companies thoughtfully embed AI into their internal workflows by applying UX principles at the systems level.
That might look like:
Auditing existing processes to find places where AI can meaningfully assist
Prototyping internal tools and interactions using Gemini
Designing trust-centered workflows with transparency and control built in
Delivering playbooks or training to help teams get comfortable using AI day to day
It’s not a polished service yet. But it’s one I’m excited to shape.
Back to School: Learning to Speak AI
Since that initial spark, I’ve been investing in learning more about the technology behind the tools. I completed AI + Design Thinking by IDEO, which gave me a strong foundation in how to approach AI with empathy, systems thinking, and iterative experimentation. I’m now working through AI for Everyone by Andrew Ng on Coursera to deepen my understanding of how AI actually works and how it can be applied responsibly across teams.
The more I learn, the more confident I feel in helping companies bridge the gap between AI hype and practical, human-centered systems design.
What’s Next
I’m still defining what this offering will become, but the opportunity is clear. More and more teams will be integrating AI into their operations, and they’ll need systems that are smart, efficient, and designed with people in mind.
