Skip to content
All projects

AI product · Concept

One place to find the right AI tool, based on what you're trying to do.

Type
Speculative product concept
Year
2025
Platform
Mobile and web
Focus
Discovery, personalisation, continuity
MindWare home and assistant screens on two phones

At a glance

Problem

AI tools are scattered, app stores rank them by popularity, and every new tool needs another account.

Solution

An AI app store with intent-based search, a conversational guide and one account across every app.

My role

Concept, flows, prototypes and visual identity. Prototypes were tested with users to refine navigation.

Outcome

A speculative concept with no shipped results, plus clear next steps for privacy and developer tooling.

Problem

Thousands of AI tools, and no good way to choose

New AI apps launch faster than anyone can track, and traditional stores surface what's trending rather than what fits your task. Each one then demands its own sign-up, often for a single use.

5,000+

AI mini-apps launched monthly in 2025, as cited in my research

Insights

What the research showed

Intent beats popularity

People want recommendations based on their tasks, history and goals.

Fragmentation kills retention

Switching between disconnected apps makes people abandon them.

Familiar is faster

Proven app store patterns lower the learning curve, so refine them rather than reinvent them.

Key decisions

How the design answers it

  1. 01

    Search by what you need

    Search understands natural language like “I need help writing a report” and filters by task, not just name.

    MindWare search screens with results
    Intent-based search.
  2. 02

    A guide, not just a catalogue

    An in-app assistant suggests, compares and explains tools in context.

    MindWare assistant conversation
    The assistant.
  3. 03

    One account for everything

    A single sign-in across the ecosystem, like one PlayStation account across its store.

    MindWare single sign-in screens
    One account, every app.
  4. 04

    Personalisation you control

    Onboarding asks for goals and preferences, and data access is permission-based and transparent.

    MindWare onboarding screens
    Onboarding.

Final design

The screens

MindWare across laptop and phone
Progress and preferences sync across devices.
MindWare brand identity
Brand identity.

Outcome

Next steps

  • Define an ethical, transparent privacy and permissions model.
  • Explore SDKs and tools so developers can publish quickly.
  • Build a design language that scales across third-party apps.

Want the full process? Read the original write-up on Medium(opens in a new tab).

Next projectKnock Knock