Groop AI - Designing Safe AI Experiences for Families

Group AI is a family-focused AI learning platform designed to provide age-appropriate AI experiences for children, teenagers, and parents through a shared but personalised product ecosystem.

As UX Lead, I helped shape the early product experience designing differentiated interfaces, learning interactions, and family workflows that made AI more usable, educational, and developmentally appropriate across age groups.

Client	Confidential Client / via Affluent Technology
Role	UX Lead
Product Type	AI Learning Platform
Users	Parents, Children (5–12), Teens (13–17)
Team	2 UX Designers
Status	Client Product in Development

Designing AI for Different Stages of Development

The biggest challenge in Group AI was not designing a chatbot interface, it was designing AI experiences that behaved appropriately for entirely different stages of cognitive and emotional development.

A six-year-old, a teenager, and a parent should not interact with AI in the same way.

Each required different expectations, interaction models, language patterns, educational support, and behavioural safeguards.

The product needed to solve several difficult challenges at once:

• making AI age-appropriate for children, teens, and adults

• supporting parental awareness without creating intrusive surveillance

• encouraging learning instead of providing shortcut answers

• designing emotionally responsible AI responses for sensitive situations

• creating safe interactions without making the experience restrictive

• balancing education, exploration, and engagement

One of the hardest design problems was defining how AI should respond when children asked emotionally sensitive or potentially harmful questions.

The experience needed to avoid harmful direct responses while instead guiding children toward confidence, support, and healthier thinking.

This required treating AI not simply as a chatbot, but as a behaviour-driven learning ecosystem.

Children don’t think:“I want structured learning.” They think:“I want to play.”

Designing for Behaviour, Not Just Age Groups

One of the most important design decisions in Group AI was recognising that age segmentation alone was not enough.

Children between 5–12 do not approach technology the same way teenagers or adults do not just because of age, but because of behaviour, motivation, and how they naturally engage with digital experiences.

Through early research and product thinking, one pattern became clear:

Children in this age group were far more likely to engage through play, exploration, voice interaction, and storytelling rather than structured learning interfaces.

This changed the design direction significantly.

Instead of presenting AI as a traditional chat tool, we created a more playful interaction model built around:

• quiz-based learning

• interactive storytelling

• game-led exploration

• voice-first conversations

• creative prompts like drawing, adventures, and imagination-based discovery

Rather than making learning feel like a task, the experience was designed to make education feel naturally engaging.

The long-term product vision extended this even further embedding learning into gameplay itself, allowing children to build skills through interaction without the experience feeling overtly instructional.

Designing Growth With the User

The teenage experience required a very different design approach.

Unlike younger children, teenagers are already familiar with mainstream AI tools and productivity platforms. If the experience felt overly simplified or childish, adoption would likely fail.

The challenge was creating an AI environment that felt capable, familiar, and genuinely useful while still remaining educational and age-appropriate.

The teen experience was designed around deeper learning workflows such as:

• breaking down difficult concepts

• research assistance

• guided explanations

• study support across subjects

• PDF-based learning and document interaction

• structured problem solving

Rather than acting as a shortcut answer machine, the AI was designed to support understanding and learning.

One particularly important design challenge was transition.

As younger users grow, their expectations change quickly. Moving them abruptly from a playful child interface into a mature productivity environment would create friction.

Instead, we explored a progressive transition model gradually introducing more advanced capabilities and interface patterns over time, allowing the product to evolve naturally alongside the user.

This approach reduced behavioural disruption and supported long-term product adoption.

Great products grow with their users instead of forcing users to relearn behaviour overnight.

Balancing Independence With Parental Guidance

Designing the parent experience required a very different mindset.

Parents needed a product that supported their responsibilities without creating unnecessary monitoring behaviour or adding more digital stress to already busy routines.

The goal was not to create a surveillance product. It was to create supportive visibility.

The parent dashboard was designed as a calm, minimal AI workspace where parents could manage their own tasks while still staying meaningfully connected to their children’s learning journey.

From the parent environment, users could:

• switch between child profiles through dedicated child agents

• understand broader learning interests and behavioural patterns

• ask contextual questions about how a child was engaging

• set learning interests and educational goals

• guide future learning directions

• create shared family chats for collaborative activities

• save conversations into projects for future use

Safety-sensitive situations were treated differently.

Rather than exposing every interaction, the system focused on meaningful alerts where parental support may genuinely be needed.

Examples included:

• excessive screen-time behaviour

• emotionally concerning patterns

• potential bullying discussions

• wellbeing-related warning indicators

This allowed parents to remain informed where it mattered, while still preserving independence and exploration for children.

Trust is built through meaningful visibility not constant monitoring.

Designing a Multi-Agent Family AI Ecosystem

Group AI was never intended to be a single chatbot experience. The product was designed as a connected AI ecosystem where different users interacted with purpose-built experiences aligned to their needs, behaviours, and responsibilities.

Each AI experience required distinct interaction logic.

Child AI (5 –12) Designed around play, voice interaction, storytelling, guided curiosity, and emotionally safe learning support.

Teen AI (13 –17) Built as a more capable study-focused assistant with document learning, structured explanations, research support, and deeper educational workflows.

Parent AI A calmer productivity-oriented environment designed for family coordination, learning guidance, and meaningful oversight.

Family AI Shared collaborative experiences such as trip planning, learning together, and family conversations.

Rather than treating AI as a one-size-fits-all tool, the product was designed around behavioural context and user intent.

This required thinking beyond interface design and into AI interaction strategy how responses should feel, what boundaries should exist, and how trust should be maintained across the ecosystem.

Client Validation & Product Direction

Group AI was created as an early-stage product vision to define how a family-focused AI learning ecosystem could work across children, teenagers, and parents.

The initial concept and UX direction received strong positive feedback from the client, particularly around the clarity of the experience, differentiated AI interactions, and the overall simplicity of navigating what could easily have become a highly complex product.

One of the strongest outcomes was validating that a multi-user AI ecosystem could feel approachable rather than overwhelming.

The product is currently progressing into deeper user experience refinement and development, with ongoing focus on interaction behaviour, usability, and AI experience optimisation.