Self-service where all you need to do is say what you want
Moi Mobiili operates in one of Finland's most competitive consumer markets. The company's operating model relies heavily on digital self-service: customers manage their subscriptions, billing, and services primarily through the Mun Moi app.
As the customer base grew, Moi identified the need to make customer interactions more automated and AI-assisted. The Mun Moi app already offered comprehensive self-service functionality, but Moi wanted to enrich the service experience with an AI that could truly serve customers. The decision was to go with a conversational interface where customers could simply say what they wanted to do.
The business objectives were clear: a better customer experience, higher self-service adoption, reduced customer service load, and more effective responses to customer needs. Equally important was ensuring that AI development would remain in the hands of the business, not solely the responsibility of the technical team.
How does the Moi AI assistant work?
The Moi AI assistant works as a conversational layer on top of the app's existing functionality. Customers can, for example, check outstanding invoices, retrieve their PUK code, or make changes to their subscription, simply by saying what they want, either by typing or speaking.
The AI was designed as an integral part of the existing service, not a separate channel. This keeps the user experience consistent and allows the service to evolve in a controlled way.
The solution was built using Flutter, Google Cloud and Firebase ecosystem services, and Codemate's Rebel AI Studio platform, enabling Moi to develop and test the assistant's capabilities quickly, without lengthy release processes.
Business ownership — not just the technical team
Rebel AI Studio gives Moi's business and customer service teams a direct role in developing the AI assistant. New approaches can be tested in a controlled way before broader rollout, and the assistant can be continuously refined based on customer feedback and usage data.
"Codemate helped us identify how AI can genuinely improve the self-service experience and open up new opportunities in customer interactions," says Moi Mobiili CTO Kalle Vuoristo.
"What we found particularly compelling about the solution was that AI development doesn't remain a purely technical project. The business can own the development of the service too. This is the first step on a shared journey towards AI-powered customer interactions," Kalle continues.
Under the hood: how the solution is built
The Moi AI assistant is not a standalone chatbot service but a new layer built on top of the existing Mun Moi app. The solution draws directly on Moi’s backend systems, customer data, and business logic. Nothing was built alongside or separately.
Technically, this was implemented using an A2UI-based architecture (agent-to-UI), where a Gemini-based AI dynamically drives the user interface based on the user's intent. Firebase Remote Config enables staged testing of changes without a full app release. Rebel AI Studio gives Moi's teams direct access to develop the assistant independently, as part of their normal service development work.
Results: what changed for customers and the business
The Moi AI assistant has brought AI into Moi’s digital customer experience in a way that serves both the business and its customers. The AI guides users directly to the right place, making self-service smoother than ever.
"AI enables an even smoother and more helpful experience for our customers. When AI takes care of guidance and the most common tasks can be handled independently, we're also able to develop customer service more efficiently and direct resources to situations where human support adds the most value," says Petri Lahtinen, CEO of Moi Mobiili.
Business impact
• Conversational interactions lower the barrier to self-service adoption
• The solution identifies friction points in the customer experience in real time
• Usage data provides new visibility into customers’ real needs and intentions
• The service is continuously developed based on feedback and data
• Customer service resources are freed up for situations where human support adds the most value
