
Arun SasiSenior UX Analyst

Design at Reflections is becoming more human and more intentional. It is no longer just about UI aesthetics or business goals. Our designers bridge system logic and real user needs, turning what technology can do into experiences people understand and value. Through UX design and human-centered design, we transform complexity into clarity that truly works.
At Reflections, we watch the tectonic plates of interaction design shift beneath our feet. The patterns we spent decades refining are cracking open. For years, software meant clicking through menus and buttons to make things happen. We thought conversational AI would change that, but what we got instead was just a different flavor of the same limitation. We swapped buttons for text boxes.
The ground is moving toward something we call Intent-Based Design. The old model asked users to translate their goals into system language, to break down what they needed into searchable terms. The new model flips this. Systems now orchestrate the journey themselves. They do not just point you toward a destination. They walk the path for you.
But this shift comes with a trade-off. When the process happens behind a curtain, the human role transforms. We are no longer the ones doing the work. We have become supervisors checking whether the machine did it right. And the chatbox, the interface we have leaned on to bridge this gap, is starting to crack under the weight.
The chatbox was a useful stepping stone for getting people comfortable with AI. Everyone knows how to send a message. No training required. But when you need to get serious work done, the chatbox becomes a constraint. It obscures the tools you actually need and forces you to guess at the right words, hoping the system will interpret your intent correctly.
We think about this distinction a lot in our work. A window is something you glance through. You check the weather. You skip to the next song. You look up a fact. A room is where you live. It is Figma. It is Photoshop. It is where you do the actual work. Right now, most AI interactions are trying to cram room-sized work into window-sized interfaces. It does not scale.
The next phase is not about refining the chat bubble. It is about Generative UI, interfaces that construct themselves in real time based on what you are trying to accomplish. The system hands you exactly the tool you need at exactly the moment you need it. The interface becomes fluid instead of fixed.
This is where things get strange for those of us who have spent our careers designing for people. In this emerging model, the user is not always human. Sometimes it is an AI agent working on someone’s behalf. That means Invisible UX, the design of APIs and data structures, now carries the same weight as the pixels people see on their screens.
A person can look past a typo. An AI stumbles and stops. We have to design for machine legibility now. If the agent cannot parse the structure of the room it is navigating, it is going to trip over the furniture. Systems need to be built so that agents can trace their own reasoning and surface that logic back to us when we ask why something happened the way it did.
Autonomy does not mean handing the car keys to the machine and walking away. It is more like a conversation about who does what. We talk about keeping humans in the loop, but what actually works better is something we think of as co-tasking.
Picture an AI booking a flight. It handles the search, filters the options, and enters passenger details. Then it hits the seat selection map and gets stuck. Instead of guessing or giving up, it should hand that specific screen back to the person. They pick the seat. Then they hand control back to the agent to finish checkout. It is collaborative, not a complete handoff.
We also need to stop designing as if AI is going to work flawlessly. Things break. Most AI failures come from not planning for confusion. We have not built escape routes when the system gets lost.
What we need is a Resilience Layer. This means anticipating where an agent is likely to fail and making sure that when one piece breaks, the whole system does not collapse. Trust is not built on perfection, but on recovery. When humans and AI collaborate to solve an error together, they solve it more than three times faster than either one working alone.
Our role as designers is shifting. We stand between the logic of machines and the needs of the people who use them, translating capability into meaning. It is no longer just about business goals or visual polish. Because if we overlook the ethics of what we build, we are not just shipping flawed software, we are quietly shaping the kind of future people will have to live in.
Author: Arun Sasi, Senior UX Analyst