AI-Powered Customer Loyalty Platform
One of the most valuable things software engineers can experience is building a product from scratch.
Not maintaining legacy code.
Not implementing isolated tickets.
But actually taking an idea from:
- concept
- to architecture
- to UX
- to implementation
- to deployment
…while collaborating remotely with a real engineering team.
That’s exactly what happened during another edition of The Startup Experience at DevAccelerator.
This time, the team — Diego, Noemí, and Patsy — built LoyalT, a customer loyalty platform designed to help small businesses modernise how they retain and engage customers.
And what stood out immediately was how practical and scalable the idea actually was.
The Problem Behind LoyalT
Most small businesses still rely on physical loyalty cards.
You know the ones:
- coffee shop stamp cards
- paper rewards systems
- plastic membership cards
The problem?
Customers lose them.
Businesses get no real analytics.
And the experience feels outdated.
The team wanted to solve this by creating a digital-first loyalty ecosystem.
Their idea was simple:
Allow businesses to create digital loyalty cards connected directly to Google Wallet, while also giving business owners access to customer engagement analytics and campaign insights.
Instead of just replacing physical cards…
they wanted to create an intelligent loyalty infrastructure.
Why The Product Was Interesting
What made LoyalT different from many existing loyalty platforms was the combination of:
- Digital wallet integration
- Dual dashboard architecture
- Business analytics
- Future AI-driven customer segmentation
The product had:
- A dashboard for businesses
- A separate experience for end users/customers
This meant businesses could:
- register their company
- configure rewards programs
- manage campaigns
- analyse customer behaviour
While customers could:
- store loyalty cards digitally
- accumulate points
- interact seamlessly through Google Wallet
The team also discussed future AI integrations, where the platform could eventually:
- generate weekly customer reports
- identify customer behaviour patterns
- create segmented marketing campaigns automatically using AI insights
That’s where the idea became much bigger than a simple loyalty app.
It started moving closer to a lightweight CRM + analytics platform for SMBs.
They've created a great community where there are people in the same situation as you, and we learn from each other.
Choosing the Right Features First
One of the strongest parts of the project was how pragmatically the team prioritised features.
Instead of trying to build everything…
they focused first on the core value proposition.
The priorities were:
- Google Wallet integration
- seamless user onboarding
- loyalty point management
- business statistics dashboards
The idea was:
if these pieces worked well, the product already delivered meaningful value.
That level of prioritisation is something many early-stage teams struggle with.
The Engineering Stack
The frontend was built using:
- React
- Vite
- ShadCN UI
Pachi led much of the frontend development and UX thinking, focusing heavily on creating intuitive user journeys for both businesses and customers.
The backend stack included:
- Java
- Spring Boot
- Supabase
Noemí worked extensively on:
- API integrations
- authentication/security
- database connectivity
- Google Wallet integration architecture
Meanwhile, Diego acted as the technical coordinator of the project, helping organise:
- GitHub project management
- feature ownership
- ticketing
- sprint planning
- technical direction
Real Product Thinking (Not Just Coding)
One thing I found particularly interesting was how much attention the team gave to UX and product behaviour.
Pachi explained that she researched how loyalty systems currently work in Ecuador and analysed the user flows of existing businesses already using similar systems.
Instead of inventing random flows…
they studied real customer behaviour first.
She then designed the experience in Figma before implementation, ensuring alignment across the team.
That’s a very real-world product workflow.
The Biggest Technical Challenges
Each engineer faced different challenges during development.
For Diego:
- working across unfamiliar backend technologies
- supporting multiple areas simultaneously
- coordinating the team technically
For Noemí:
- integrating external APIs
- handling Google Wallet configuration
- designing secure backend architecture
- implementing authentication flows using Spring Boot
For Pachi:
- designing scalable UX flows
- balancing business/admin experiences with customer simplicity
- translating ideas into intuitive interfaces
And like every real engineering team…
they also faced coordination challenges:
- overlapping work
- duplicated features
- scheduling conflicts
- balancing development with full-time jobs
Which honestly made the experience even more realistic.
Using AI During Development
Another fascinating aspect of the project was how naturally the team integrated AI into their workflows.
This wasn’t “AI replacing developers”.
It was developers using AI to accelerate execution.
Diego used:
- ChatGPT
- OpenCode
- AI-assisted planning and code generation
Pachi used AI heavily for:
- frontend scaffolding
- reusable component generation
- UI acceleration
- iterative code refinement
Meanwhile, Noemí used ChatGPT more as:
- a learning assistant
- a way to reinforce Java and Spring Boot concepts
- support for backend implementation decisions
This distinction is important.
The best engineers today are not simply “using AI”.
They’re learning how to collaborate with AI effectively depending on the task.
What They Learned Beyond Their Jobs
One thing that came up repeatedly during the interview was this:
In many corporate jobs, engineers rarely get to think like product builders.
Requirements already exist.
Architectures already exist.
Roadmaps already exist.
But here?
They had to:
- define the product
- prioritise features
- think about future scalability
- own technical decisions
- collaborate cross-functionally
And that changes how engineers think completely.
Final Thoughts
What I loved most about this team was how pragmatic their approach was.
No overengineering.
No unnecessary complexity.
No trying to impress with fancy buzzwords.
Just:
- clear product thinking
- good collaboration
- practical engineering
- and smart use of AI tooling
Exactly the kind of mindset modern software teams need today.
As AI continues changing software engineering, the engineers who will stand out are not the ones who simply code fastest.
They’ll be the ones who can:
- think like product builders
- collaborate effectively
- adapt quickly
- and leverage AI intelligently
And this project reflected all of that.
Results that speak for themselves
Engineers
Months
Product
