A self-initiated concept exploring the hardest UX problem in AI apps: what do you show when the chat is empty? I designed a multi-tool AI assistant — chat, image generation, summaries and prompt packs — built around the idea that nobody should have to stare at a blinking cursor.
What this shows
AI-product UX: beating the blank-box problem — designing the empty state of an AI assistant so non-prompters create something on their very first try.
What I designed for
The challenge
The market was flooded with ChatGPT clones: a cursor, an empty input, and a silent “…now what?” Powerful underneath, paralysing on top. Most people don’t arrive knowing the perfect prompt — they arrive with a vague need and bounce when faced with a blinking line.
My brief: turn raw AI power into something a first-timer can do something with in ten seconds — without ever staring at an empty box.
What it does
Ask anything, in plain language.
Generate art from a description.
Web pages & PDFs, distilled.
Specialised helpers for real tasks.
Rewrite, translate, and more.
Write & explain code.
Discovery
I don’t know what to type
The blank box is a wall. Without a starting point, people open the app, stall, and close it.
What can this even do?
Capabilities were invisible. Users had no idea image gen, summaries or assistants existed.
Why would I pay?
Without feeling the value first, the paywall landed as a tax — not an upgrade.
From a teardown of how mainstream AI assistant apps handle the first session, and the recurring “I don’t know what to type” complaint in their public reviews — framed as the hypotheses a live product would test.
Key decisions
I traced the blank-box problem to its source, made three structural bets to solve it, and designed each in detail. These are the decisions I'd stand behind in any product critique.
The hardest problem in AI design isn’t the model. It’s the empty input field.
What the concept demonstrates
Opening to vivid tool cards — not a cursor — gives every new user a ten-second starting point without requiring any prompt knowledge. The bet: discoverability drives first-session creation.
Specialised assistants and one-tap prompt packs remove the skill barrier. Users get high-quality results without ever writing a prompt — the bet is that this drives repeat use.
A credit system lets users generate something real before the Pro gate. The bet: "more of what you just loved" converts better than "pay to start."
What I'd validate next
The primary signal: did the tool-first home actually get more new users to produce something vs. a blank-chat baseline? I'd run an A/B test on the home screen on day one.
Which tool card drives the stickiest users? Knowing whether image gen or assistants builds the stronger habit would inform what to expand and what to promote.
Does running out of credits at a high-value moment (first generated image) correlate with upgrade? I'd instrument each credit-spend event to find the exact trigger.
Reflection
A few power users wanted to just type. I’d keep the tools but make a plain conversation one tap away from anywhere.
“+11” wasn’t self-explanatory. I’d show what a credit buys at the moment it’s spent, not just a number.
Beautiful brand, but busy behind dense chat. I’d reserve the loud gradients for moments, not every surface.
Design system
Palette & gradient
Typography — Inter display, DM Sans body, Space Mono accents
Key components
Vivid gradient entry points — the home of the app.
Clean dark bubbles over the fluid background.
The gradient action, with credit context.