Agentic AI as the front door to customer service

Helen, one of Finland’s leading energy companies, is transforming customer service with a bold, agentic AI first approach. By building a scalable network of AI agents from scratch, Helen is redefining how customers interact in digital channels, making AI the primary entry point for written customer service.

Helen_logo

50+

Agents deployed across channels within 12 months

4.6/5

Customer Effort Score for relocation

3 days

To deploy a new AI agent

Watch Helen's success story!

The challenge

Helen faced growing pressure to provide instant, always-available customer service for a wide range of energy-related topics: contracts, billing, relocation, grid services and more... Customers often reached out outside service hours or struggled to find the right digital path, leading to unnecessary contacts and repeated questions.

To fix this, Helen had to start completely from scratch. There was no existing AI knowledge base, no conversational structure, and no previous chatbot to build on. Every use case had to be defined from the ground up but instead of introducing isolated solutions, Helen built a single, scalable agentic AI foundation.

The challenge was to handle all written customer interactions while improving the experience and reducing effort for both customers and service teams - in a way that could scale across business units and use cases.

Helen Customer Story Challenge
Katja Makela Helen Finland
“Starting from scratch gave us the opportunity to design an AI-first experience, where digital service begins with AI, not as an add-on.”
Katja Makela
AI Product Owner

The solution

Helen implemented a fully NiCE agentic AI orchestration architecture, replacing traditional chatbot and NLU approaches. A growing network of 50+ AI agents is deployed across Helen’s app, customer portal and website.

All agents operate within their specific context but are connected through a shared network and are centrally coordinated. This ensures consistent routing, context handling, and a unified customer experience across channels. Some of the specialized agents at Helen include:

  • Customer support agent that answer questions and guide users to the right self-service options
  • Service navigation agent that direct customers to the most relevant digital journey or action
  • Solution agent that complete end-to-end tasks, such as the relocation agent that moves an energy contract to a new address
  • Relocation agent that handles address changes autonomously.

The AI agents resolve requests, and enable seamless escalation to service teams when needed. The introduction of solution agents marks a shift from simple guidance to full task execution within a single interaction.

 

Helen Story Child

The impact

Helen has established a fully agentic AI customer experience within a year, transforming digital interactions across channels. AI agents now act as the central entry point for written customer service, providing consistent guidance, reducing friction, and enabling efficient self-service.

The solution delivers high-quality interactions at scale, with over 80% positive customer feedback.

Automation has reduced manual workload across customer-facing and internal processes while accelerating service delivery. Solution agents mark a key advancement, with the relocation agent achieving a CES score of 4.7 and a 100% resolution rate, demonstrating end-to-end process automation.

The agentic AI approach enables rapid scaling, with development cycles reduced to a structured two-week model and new agents deployable within days.

Helen has evolved from a greenfield setup to a multi-agent architecture that enables continuous expansion, and long-term transformation.

Text message White Icon

Seamless cross-channel experience

 

 

Success White Icon

80% positive customer feedback

 

 

Technology 02 White Icon

Multi-agent architecture

 

 

Want the same success with NiCE Cognigy?

Use AI to deliver better service and reduce costs, while supporting customers and service teams. Around the clock, at scale.