Dalkia adopts calldesk's conversational AI to absorb call spikes


Claude Desmoucelles

Head of Shared Operational Services

Voice agent project overview


Call center agents, on 2 sites


Departments for experimentation

500 000

Calls Automated


Customer service availability

Générale de Téléactivités, Dalkia’s contact centre in France

« My name is Claude Desmoucelles, and I am the managing director of the Générale de Téléactivités, a medium-sized business that acts as Dalkia’s Customer Relationship Centre for France. I am also responsible for Shared Operational Services within Dalkia’s Operations Department.

Dalkia, an EDF Group subsidiary with 16,000 employees, is one of France’s leading energy services companies. We promote the use of local renewable energies (biomass, geothermal energy, heat recovery from industrial processes, etc.) and enable our customers (healthcare institutions, housing, office buildings, public buildings, etc.) to reduce their carbon footprint by lowering their energy bills. All this, with equal comfort.

At Générale de Téléactivités, most of our work is devoted to Dalkia. Our contact centre is spread over two sites, in Rouen and Bordeaux. We employ the equivalent of 120 FTEs and operate 24 hours a day, 7 days a week. During call spikes periods, we recruit temporary staff to reinforce our workforce. We also use other service providers. »

The problem: high call spikes during the winter season

« Since Dalkia’s activities are focused on energy, our contact centre receives very large volumes of demand each year starting in October, when the facilities are “re-ignited”. This very high number of calls is difficult to absorb, despite staff reinforcements! The result is a waiting time that I consider too long for customers, as it could generate dissatisfaction.

In addition, it could encourage customers to call several times, and lead to an increase in the number of complaints. »

The solution: a conversational AI to orient customers and automate information retrieval

« To avoid this, in 2018 we conducted an audit on the optimization of our operations and customer satisfaction, in order to improve our processes and better absorb these call spikes. We quickly identified conversational AI -a voicebot applied to the telephone channel- as a relevant solution, which would allow us to automate some of the conversations and free up the time of call agents to better manage calls.

After a tender process that allowed us to choose the calldesk technology, we analyzed the conversations to identify a list of repetitive questions.

We then programmed the conversational AI to respond directly to incoming calls from individuals, without waiting time. A natual language IVR (Interactive Voice Response) is capable of directing each caller in the right direction by asking questions to qualify their request.

The conversational AI then retrieves information and asks the caller all the simple questions we have identified with our teams. When an advisor from our subsidiary retrieves the call, the CRM tool then allows him to access this information, saving him precious time. »

Our feedback: a successful test, and a generalization in progress throughout the country

« In order to support the change and the implementation of the new solution, we first tested the conversational AI on calls from two departments, which had volunteered. This was in April 2019. Then we increased the test volume in July 2019, including 9 departments, spread over different regions. Very early on, we wanted to involve both management and employee representative bodies, which enabled the project to move forward on schedule.

Since January 1st, 2020, we have moved into a generalization phase, so that the conversational AI can now absorb all regions and calls. This represents approximately 500,000 automated calls each year. calldesk teams have accompanied us to ensure this increase in workload. We found a good way to work together.

This gradual generalisation should enable us to confirm the project’s ROI, thanks to the reduction in the average length of conversation between our employees and the caller. By freeing up our colleagues’ time, it also allows us to reduce waiting times, and therefore improve the satisfaction of our end customers. Thus, all parties should benefit from the implementation of this new initiative! »

Automate repetitive calls in your contact center

Contact us to schedule a personalized demo