Which Company's Case Study Is This?
Vodafone is one of Europe's largest mobile and fixed network operators. With hundreds of millions of customers and numerous employees worldwide, its HR and customer support departments received an immense volume of inquiries daily.
The Challenge Vodafone Wanted to Solve
Employees and customers had to wait to connect with a support representative to get answers to simple questions such as vacation requests, checking billing plans, or various procedural inquiries. Meanwhile, support representatives were spending a significant amount of time on repetitive inquiries, facing the challenge of not being able to focus on more complex, value-added tasks. The company aimed to improve these experiences using digital technology, providing information quickly and efficiently.
How AI Was Used
Vodafone introduced TOBi, an AI-powered virtual assistant. TOBi initially started as a text-based chatbot, built utilizing IBM's AI platform, 'IBM Watson Assistant,' among others. Through Natural Language Processing (NLP) technology, it understands the intent and tone of questions phrased in human-like language and provides answers in a conversational format. TOBi not only answers simple FAQs but can also perform specific actions, such as changing customer account information when necessary. It also features a seamless handover function to a human agent if an issue cannot be resolved.
Estimated Architecture
The following is an estimated architecture based on public information and general configurations. It does not definitively represent the actual internal structure.
- ▸Input Data: Natural language question text entered by customers and employees via chat screens on websites, official apps, WhatsApp, etc.
- ▸AI Processing/Model or Search Layer: Conversational AI platforms like IBM watsonx Assistant interpret user intent using Natural Language Understanding (NLU) technology. It searches a pre-trained knowledge base of internal regulations, product information, etc., to generate optimal responses. Generative AI technologies like IBM watsonx.ai may also be utilized to improve dialogue quality and automate testing.
- ▸Output to Business Systems: Display the generated response text in the chat interface. If procedures such as address changes or plan changes are required, it performs processing via API integration with backend Customer Relationship Management (CRM) or Human Resources Information Systems (HRIS), or guides users to the relevant page.
- ▸Monitoring and Governance: Monitor dialogue count, resolution rate, customer satisfaction (NPS), etc., using a management dashboard. Analyze unresolved questions and dialogue logs to build a closed-loop feedback system for continuous AI knowledge updates and improvements.
Implementation Effects and Key Takeaways
- ▸Streamlined Inquiry Handling: TOBi handles millions of inquiries annually, resolving many of them on first contact. This has allowed human agents to focus on more complex issues.
- ▸Improved Customer and Employee Experience: Users can now get immediate answers 24/7 without waiting. This has reportedly led to a significant increase in customer satisfaction (NPS).
- ▸Cost Reduction and Productivity Improvement: Automation of routine tasks has contributed to reducing operational costs, including personnel expenses for support departments.
- ▸Key Takeaway: This initiative is not merely automation for cost reduction. The important point is that AI is positioned as a 'critical team member,' and by fostering collaboration between humans and AI, it improves the experience for both customers and employees (CX/EX).
What Japanese Companies Can Learn
Vodafone's case study offers valuable insights not only for large corporations but also for many small and medium-sized enterprises. Departments with numerous routine inquiries internally, such as HR, general affairs, and IT help desks, are strong candidates for AI chatbot implementation. A realistic approach is to start small with FAQ support in a specific area, gradually expanding the scope of support while accumulating conversational data. This enables companies to achieve both operational efficiency and improved employee satisfaction.