WHO?
I was responsible for strategy, research, synthesis, and overall design of the MVP. I worked with the founder and developer in tight feedback loops, aligning business goals, user needs, and technical feasibility to bring clarity and structure to the MVP.
WHAT?
AI Aggregator platform for Mobile & Web - www.internet.io
WHY?
Everyday users find themselves overwhelmed by so many AI models, each offering different capabilities, personalities, and limitations, and they often don’t know where to start. Most AI platforms today are either too technical, too open-ended, or designed for power users. We set out to answer a critical question:
"How might we make AI feel more approachable, useful, and familiar by creating an all-in-one platform that reduces technical complexity?"
Looking at the AI Landscape
The concept had early interest and momentum but lacked clarity and direction. I joined early to define the brand, establish the UX and product structure, and set a clear design direction.
Before exploring how it might take shape, I wanted a clear understanding of the AI landscape, user pain points, and gaps.
I created a 20-question survey covering demographics, AI usage patterns, feature preferences, and frustrations. We distributed it via Prolific and received 485 responses.
5 KEY PAIN POINTS FROM SURVEY RESPONSES
Trust Gaps
How do I know this answer is correct?” came up very often.
Technical Interfaces
Most AI tools are designed for power users, not casual searchers. Users found AI tools “too technical” or “confusing”.
Desire to Compare
91% preferred a single platform that aggregates responses from multiple models.
Organizational Needs
Users want to save, label, and revisit answers not just generate once and forget.
Tool Overload
89% of users said they switch between 2–5 AI tools regularly.Too many tools, too many tabs.
I identified popular search engines, AI aggregators, and LLM platforms, and studied their approach to simplicity, user-friendliness, design patterns and the overall learning curve.
opportunitIES Identified from research
Create a search-first experience that feels familiar, reduces cognitive load and is easy to use.
Aggregate trusted AI models in one space to compare efficiently without switching tabs.
Make answers saveable by organising them into folders for easy access.
Mapping out the user flow
I started by mapping out the core interaction. A user asks a question, sees responses from multiple AI models, and chooses one to continue the conversation with. This flow helped me validate that the experience could stay intuitive even as features like chat, saving, and folders were layered in.
I didn’t want users to feel lost or overwhelmed. The goal was simple: ask a question, explore answers, and decide what feels right, all in one clean flow.
Starting with Low Fidelity MVP
At the low-fidelity stage, I focused on shaping the experience around one core question:
“What’s the most important thing the user needs to see and do — right now?”
This helped me structure each screen to reflect the product’s value clearly and instantly.
My focus here was on:
Clear visual hierarchy to help conversion
One clear task per screen to reduce cognitive load
Simple, clutter free, and familiar layouts using well-known patterns
Encouraging guided exploration without overwhelming the user
Simplifying the Path to Answers
Next, I had to fine-tune the flow. The challenge was to make sure every step felt intuitive, from how they compare and interact with them to how users land on answers.
#1 Getting preview & comparison right
To shape a clear and usable comparison flow, I tested three layout directions. Each explored different ways to present responses, highlight models, and support interaction.
I decided to combine the strengths of Layout 1 and Layout 2 by adding a toggle button to switch between a minimised and maximised preview. This allowed users to choose how they prefer to explore responses, whether they want to quickly scan multiple answers or focus on one in detail.
#2 Choosing the Right Flow After Search
Based on my initial user flow and low-fidelity wireframes, the plan was to show a list view after search to let users scan and choose a response. I later explored an alternate path that skipped the list view and opened the first answer directly, reducing interaction cost and giving immediate access to a full response.
#3 Aligning with Stakeholders & Validation
Once I had a clickable prototype, I shared two flow options with the founder and developer. We discussed the pros and cons of each approach, focusing on clarity, effort, and how well they supported the product’s core value. This helped align the team and validate the direction before moving further into chat and folder features.
Stakeholder Feedback
Chose the First Flow for better focus, readability, and a cleaner layout. Unlike Flow 2, the first answer stays closed by default, giving users control over what to view.
Removed the AI model list on Home to reduce cognitive load and avoid issues as the list grows over time. This was a key concern from the founder
Added text labels to Copy, Regenerate, and Save for better clarity, instead of relying only on icons.
Discussed the scope for adding AI Agent store down the line allowing users to create & use AI Agents for any use.
Move forward with this flow and design for chats, folders, and mobile-first use for signed-up users and Free users.
Moving from MVP to a Scalable Product
After shipping the core flow, I moved into shaping what comes next. I focused on defining scope, prioritising features, and making sure we stayed clear on what matters.
This phase was about setting the product up to scale, keeping it mobile first, and building towards a public launch on Product Hunt. Every decision was about growing the product without adding noise, and making sure the experience stayed simple and clear.
Prioritised Features for Launch
Consistent Chat Layout
Make the experience predictable and reduce learning curve.
Free vs signed-up user logic
Drive signups while letting users try the core value before committing.
Folder organisation for chats
Help users easily revisit important conversations without clutter.
Mobile-First Optimisation
Ensure smooth use across all devices, since most users come from mobile.
The core challenge was integrating new features without disrupting the existing flow while leaving room for future expansion.
The solution was a fixed sidebar that brought search, chat, and AI agents into one place.
It made navigation smoother, worked well on mobile, and set up the layout to scale as the product evolved. This also made it easier to implement the free vs signed-up user logic by letting us hide core features in the sidebar for free users.
It also allowed me to add the planned features for the product launch easily. Chat worked like messaging apps but was tailored for AI, and folders followed familiar design patterns from Google Drive, OneDrive, and File Explorer. This reduced the learning curve and made navigation easier.
Optimizing for Mobile
Staying responsive is built into my design process. I use Figma with Auto Layout and components to make sure screens adapt seamlessly across devices. This made mobile optimisation faster and smoother once the core flow was locked in.
Launched on Product Hunt!
Internet.io launched on Product Hunt and ranked #10 on launch day, gaining over 100 user sign ups within the first 24 hours. The response was clear. Users understood what the product did as soon as they landed on the homepage. That was the goal from the start. Keep it simple. Keep it intuitive. Make it useful from the first click.
What I Learned & What’s Next
The hardest part wasn’t designing features, but knowing when to stop. Keeping the experience focused while making room for growth required restraint, clarity, and alignment across the team.
Key Takeaways
Designing for clarity always pays off
Auto Layout and Figma components helped scale across screens fast
Small UX decisions made early saved time later