How a European MVNO Decreased Customer Service Costs & Increased NPS

Key metrics

The following data are averages of achievements across four regions in Europe:

  • 7% increase in chatbot resolutions over one year.
  • NPS scores increased by an average of 35 points in one year.
  • Regions with negative NPS saw an increase of over 40 NPS points, taking the overall score from negative to positive.
  • An average 10% decrease in customer service costs in one year, despite an over 20% increase in the average customer base.

Overview

A large MVNO in Europe operating in multiple countries is undergoing rapid digital transformation. As part of this initiative, the telco explored ways to improve its contact center efficiency without compromising customer experiences. This aligns with the enterprise’s ultimate mission; to deliver affordable, flexible plans without compromising customer experiences or service quality.

Achieving this goal meant deeply understanding the customer journeys in the contact center and its operations (including the role of human agents). The MVNO chose to meet these goals by leveraging a conversational AI solution to enhance self-service experiences. 

Engaging

M.M. offers various service models—including consultation, resource augmentation, business process outsourcing, and Integration as a Service—to support enterprises in their digital transformation journeys. The international European MVNO entered into a resource augmentation partnership with M.M., onboarding both developers and a solution architect to prime the conversational AI solution to deliver in the MVNO’s regional contact centers. M.M. has been in the field of conversational AI chatbots since 2008 and our expertise includes our own conversational AI platform. 

Their vision evolved the conversational AI solution beyond functioning as a deflection tool; the goal was to have a chatbot that successfully serves customers in select journeys from end to end. 

Automating the contact centers: preliminary considerations

The North Star metric – Cost per serve  

Also known as cost per contact, cost per serve is the average cost of handling each customer service interaction. Longer interactions mean a higher cost per serve. The European MVNO aimed to decrease its cost per serve in each of its regional contact centers by resolving simple queries with deterministic chatbots.  

But, while cost reduction was definitely a key driver, this metric was selected as the North Star for additional reasons. The factors contributing to a reduction in cost per serve, and the MVNO’s approach to achieving this goal, reflect fundamental efficiency improvements in the contact center.  

For the telco in question, the journey to reduce cost per serve also facilitated reductions in average handling times, improvements in NPS and contact center efficiencies, and a thoughtful redesign of the human agents’ workflows.

Anticipating and preparing for fears around automation

A challenge that’s not unique to MVNOs is the fears employees experience when an enterprise considers widescale automation. Questions such as “Will I work myself out of my job?” arise inevitably, creating resistance to change within the organization. Thus, it’s wise for enterprises to accept this reality and prepare to manage change strategically at every level. 

The European MVNO we worked with had introduced various digital transformation initiatives prior to the contact center automation one. From previous experiences, the enterprise learned to navigate these fears by engaging employees in the planning process and conveying:

  • The “why” behind the initiative. In this case, why was the enterprise adopting contact center automation? The answer is to improve contact center efficiencies. 
  • The initiative’s expected impact. Automation can either reduce headcount or simply augment processes, helping humans increase their productivity. In this case, the enterprise considered both these outcomes; while many queries would still require a human agent to resolve, the initiative would impact the agent-to-query ratio. This could mean a reduction in headcount. 
     
    However, the MVNO is also growing rapidly, so the impact of automation could also mean existing customer support team can manage the increasing workload without increasing headcounts. 

Importantly, the international European MVNO had already cultivated a culture of digital transformation and innovation across its regional offices. Thus, its workforce was more receptive to change and prepared to adapt quickly. 

Securing internal buy-in: negotiating with stakeholders

The contact center automation initiative was driven by the European MVNO’s business unit, motivated by the desire to improve the center’s efficiency. However, technology adoption naturally requires support from the IT unit – which, in this case, was already engaged in an extensive, broader digital transformation project.  

This created discourse around which initiatives to prioritize; while IT leaned towards meeting its existing projects’ timelines, the business unit was convinced of the need to rapidly develop its chatbots capabilities.  

Convinced of the value of automation through a conversational AI solution, the business unit began exploring options to minimize its requirements from the IT unit by exploring low-code and no-code options. It eventually selected OneReach AI, a no-code platform for building AI agents for various business functions, including customer service. OneReach was selected for its potential to build and deploy customer service AI agents in the contact center, to enhance the European MVNO’s self-service offering. 

The project was implemented in a novel fashion. By using OneReach AI and M.M.’s solution architect as a full-time resource to develop the solution, the IT department was used to facilitate and even implement the integrations, providing for seamless interaction between the development team and the supporting IT team.  

Implementing the automated workflows: a stage-wise approach

The business unit had ambitious plans for chatbot automation, outlining the project’s long-term scope with projected metrics to achieve (mainly, a desired decrease in the average cost per serve). They considered three primary factors when determining which use cases to prioritize: 

  • Existing automations on the MVNO’s website and app. Automated self-service experiences offered through the telco’s website and app were quickly prioritized and replicated in the chatbots.  
  • Business value. The projected business value of the planned use cases was estimated primarily in the context of the impact on the average cost per serve.  
  • Technical complexity. The technical complexity of implementing the respective use cases was considered, as greater complexity significantly impacted the timelines for implementation. 

After implementing automations that existed on other channels in the chatbots, we prioritized use case implementation by weighing the respective business value and technical complexity. Use cases with high projected business value and low technical complexity were prioritized. 

The team’s service leaders didn’t restrict themselves to an either/or scenario, where a process was selected as either a candidate for automation or for human agents to handle independently. Several scenarios involved both chatbots and human agents, increasing the process’s efficiency while serving customers in the way they prefer. This illustrated a mature approach to the deployment of AI; AI was leveraged to augment the human agent’s workflow by performing tasks that machines can do naturally and quickly, consequently enhancing the human agent’s performance. 

For example, data protection regulations require contact centers to authenticate users before they can request certain actions. This process was entirely automated by a chatbot, saving agents 30-100 seconds and reducing the average handling time by progressing customers more quickly along the queue. And on the other hand, where the customer requested support from a human, it was promptly provided. 

Measurement and impact

The European MVNO uses monthly and quarterly reporting to assess both the chatbots’ performance in silo and the wider impact on the contact centers’ efficiency. This assessment is performed for the enterprise’s four regional contact centers. Some of the key metrics assessed include: 

  • Chatbot-centric metrics: 
    • Engagement with bot. What percentage of customers choose to engage the chatbot? 
    • Resolution rate. What percentage of queries does the chatbot successfully resolve? 
      NPS. How satisfied are users with their chatbot experience?
  • Metrics to assess chatbot’s impact: 
    • Contacts saved. What percentage of customer queries are resolved entirely by the chatbot? 
    • Cost per serve. Is the contact center achieving its objective of reducing cost per serve? 
    • Average handling time and wait times. Are these metrics improving as a consequence of the chatbot? 

How the customers responded to the chatbot, and their behavior are also key considerations for the contact center leaders. Thus, the team regularly tracks customer interaction journeys to observe where customers choose to engage the chatbot, and which processes they prefer to engage an agent for. They also analyzed processes where users chose to initially engage a chatbot and then requested handover to an agent. 

Importantly, the enterprise’s objective was not to move every user to the chatbots. Service leaders committed to understanding customer journeys and tailor experiences to match consumer expectations. This meant providing the option for self-service chatbot experiences at touchpoints that customers wanted to engage with it. Metrics such as NPS are used to gauge customer satisfaction to help service leaders verify how consumers feel towards the automated experiences. 

Scroll to Top