Deep-dive

Take a closer look at what went into delivering this solution.

Discovery

The initial discovery phase began with a simple but pervasive problem: people routinely forget or struggle to communicate essential information about their physical environments. Whether it’s a family member navigating daily routines at home, a caregiver trying to orient a patient in a healthcare setting, or an Airbnb host explaining how to operate the heating - the gap between physical spaces and accessible information leads to frustration, inefficiency, and even risk.

User interviews and field research confirmed that no existing mapping or signage solutions offered the level of personalisation or flexibility that modern users expect.

This led to ideation around augmented reality (AR) as a bridge between the physical and digital. We iterated multiple concepts, ranging from QR code systems to static AR overlays, but user testing consistently highlighted the need for:

• Dynamic, customizable AR maps

• Simple content creation without technical skills

• Cross-sector usability

The final Minimum Viable Product (MVP) focused on an easy-to-use mobile AR app that empowers users to overlay virtual elements, such as instructions, markers, and multimedia, onto their real-world environments.

Challenges

Several challenges emerged throughout the project:

• Balancing Simplicity and Functionality

Creating an intuitive interface that offers deep customisation without overwhelming non-technical users was a key design challenge.

• Device Compatibility

AR experiences vary significantly across devices. We had to develop robust compatibility solutions to ensure consistent performance on both high-end and lower-spec smartphones.

• Environmental Variability

Lighting conditions, space dimensions, and cluttered environments made reliable AR marker placement and retention difficult. We integrated advanced plane detection and spatial anchoring to mitigate these issues.

• User Engagement

We discovered early that some users were hesitant to engage with AR technology. Iterative user testing helped us refine onboarding processes and reduce friction.

Use of AI

AI wasn’t just a feature of the final product, it was a fundamental enabler in the development process itself.

• AI-Assisted Development with GPT

From the early stages of prototyping through to MVP completion, we leveraged GPT-based coding assistants (including GPT-3.5 and GPT-4) to accelerate the app’s development. This included generating Swift and ARKit code snippets, troubleshooting complex AR scene setups, and iterating on user interaction logic. Rather than relying solely on manual coding or outsourcing, the use of GPT allowed for rapid prototyping, reduced development costs, and the ability to test multiple approaches to spatial computing challenges quickly.

• Problem-Solving and Debugging

Throughout development, GPT was used extensively to solve roadblocks, such as gesture recognition integration, node manipulation, and ensuring compatibility across different iOS versions and devices. This dramatically reduced delays and enabled non-traditional developers to contribute to functional code segments.

• Iterative Design Support

Beyond code, GPT supported design iteration by generating ideas for user experience (UX) flows, onboarding processes, and even copywriting for in-app prompts. This allowed the team to align technical development with user-centric design principles efficiently.

• AI-Powered Future Roadmap

While the initial MVP focuses on user-generated AR content, future versions will incorporate AI to automate aspects of map creation, such as:

• Recognizing common room layouts and suggesting object placements.

• Adapting AR content dynamically based on user behavior.

• Offering intelligent prompts for accessibility and efficiency improvements.

In short: GPT-based AI didn’t just enhance MappAR - it enabled it.