Frictionless customer-first care

Eni Plenitude improved the flexibility and accuracy of its IVR call handling with Lucilla, a voicebot that uses Genesys and Google Cloud Dialogflow. Lucilla correctly identifies the customer in more than 80% of cases and successfully routes 90% of calls to the correct agent the first time. Over the last year, as a result of several initiatives including Lucilla, the company has seen a 30% reduction in agent callbacks and a 30% increase in Net Promoter Score (NPS).

30% reduction

in agent callbacks

100% of incoming calls

processed by AI

80% customer

identification accuracy

90% of calls routed

to the right agent

30% increase

in NPS

Becoming a trusted energy supplier

Plenitude is a unique energy Società Benefit that integrates production from renewables, the sale of gas and electricity, energy services and a large network of electric vehicle charging points. It serves around 10 million residential and business customers in Italy, France, Spain, Greece, Slovenia and Portugal.

Seven contact centers are at the forefront of its interactions. They operate 24/7 with 1,400 agents handling millions of phone, email and webchat contacts annually. Transforming customer experience (CX) through digital and automation technologies is a strategic goal of Eni Plenitude to provide fast, frictionless service that recognizes customers and their intent.

Eni Plenitude aims to become a trusted energy partner for retail and business customers all over Europe, thereby increasing their loyalty. The company knew success meant shifting from reactive, transaction-based communications to a more proactive CX model for personalization and value creation at scale.

Overcoming language barriers

Taking advantage of Genesys open APIs and the native integrations with Google Cloud Dialogflow, the company created a voicebot called Lucilla to improve the flexibility and accuracy of its IVR call handling via the Genesys Voice Platform. The main objective was to apply Google Cloud natural language processing (NLP) in Dialogflow to increase recognition rates and CX quality — not to only realize cost savings. The virtual assistant quickly identifies the customer and their reason for calling, then routes the call to the right team of experts to deal with the specific inquiry.

Eni Plenitude already had a relationship with Google Cloud and had good experiences with their robotic processing automation (RPA) and artificial intelligence (AI) solutions. So, Google Cloud Dialogflow was a natural choice. It also felt the NLP platform was one of best on the market and needed the least time to start the project.

However, like many innovators Eni Plenitude had to break new ground. Building a bot to understand the subtle nuances of the Italian language, like when customers pronounce meter reading numbers differently over the phone, was far from straightforward.

Results speak for themselves

The first tentative steps on that journey started almost three years ago. Since then, Eni Plenitude has been working hard with Google Cloud Dialogflow to better understand customer issues and reduce friction points.

Today, AI automation handles 100% of incoming calls, significantly improving first-call resolution (FCR).

Using Genesys and Google Cloud Dialogflow, Lucilla correctly identifies the customer in more than 80% of cases, successfully routing 90% of calls to the correct agent. Also, because the conversation feels more human-like compared to more mechanical IVR menus, there’s less propensity for the customer to hang up.

Over the last 12 months, Eni Plenitude has seen a 30% reduction in agent callbacks along with a 30% increase in NPS, as a result of several initiatives including Lucilla. Future plans include extending Lucilla to chat and WhatsApp channels, while moving to a full omnichannel CX model.

At a glance

Customer: Eni Plenitude

Industry: Energy

Location: Italy

Contact center: 1,400 agents

Challenges

  • Boost satisfaction with new customer care model
  • Increase efficiency with AI and NLP

Additional resources

PDF case study