The new colleague in customer service: a voice bot

Thanks to a voice portal, Signal Iduna has significantly improved its telephone customer service. Employees are pleased with the bot's workload, customers are pleased with reduced waiting times and low referral rates.
Signal Iduna wants to meet its customers on equal footing and attaches particular importance to the excellent quality of customer dialogue. The focus is also on telephone service. After all, this channel forms a direct point of contact between the company and its customers. The service insurer is meeting the challenge of connecting every caller to a suitable contact person with an intermediary routing instance.
As early as 2016, IP Dynamics set up a Voxtron Communication Center (VCC) at Signal Iduna on an existing Skype for Business infrastructure. At the end of 2017, the insurer decided to integrate a modern voice portal into the system. Until now, Signal Iduna worked with a simple Interactive Voice Response (IVR), which allowed callers to navigate through a voice menu at the touch of a button. However, categorization using dial keys is limited, as there are simply too few selection options.
All roads lead through the language portal
With the help of the new voice portal, the internal forwarding rate for customer calls should be reduced and waiting times for callers reduced. For this purpose, the previous IVR navigation via key selection had to be replaced by a system with natural language recognition. The new voice portal should greet callers in a friendly manner, recognize their concerns and — if necessary — authenticate them by querying their insurance number and customer data. The extensive data protection query, which previously had to be carried out during an employee interview, should be carried out automatically through the language portal in the future. After successful routing, all information that the customer has provided to the voice bot should be presented to the respective service employee at a glance via a CRM interface. In addition, a so-called “Smartlink” functionality should be integrated, which, in appropriate cases, the voice bot links to self-service offers on the company website. However, the system should under no circumstances be used to avoid employee interviews. If the query through the voice portal is unsuccessful or the caller would reject a self-service offer, the bot should communicate as directly as possible to a human colleague.
“The particular challenge of this project was to think through all possible dialogue paths in advance”, explains Dr. Jürgen Haas, who supervises the project on the part of IP Dynamics. “The prerequisite for this was to represent the entire range of formulations that customers use to describe their concerns. Finally, voice recognition must be prepared for all contingencies in order to successfully carry out the customer dialogue. ”
One and a half years from concept to go-live
The project was launched in January 2018. First, the project team developed a comprehensive concept. Formal principles were clarified, approval from the works council was obtained, data protection questions answered and determined to which goals the new language portal should route. “During the design phase, IP Dynamics' wealth of experience from comparable projects paid off”, says Andrea Schmidt, project manager on the side of Signal Iduna. “When it comes to voice recognition and dialog management, IP Dynamics has provided valuable information for fine-tuning. ”
IP Dynamics began technical implementation at the end of 2018. First, a basis was created to be able to conduct natural-language dialogues using the latest voice and customer interaction technologies. The integration of the necessary technology was achieved through the implementation of Dynamic Flow (DFW), the central workflow engine with voice portal functionality from IP Dynamics, implemented. The DFW not only enables the connection of a voice recognition engine, but also the integration of the CRM system used by Signal Iduna.
In order to take into account all potential dialogue paths and the range of phrases used by callers, keywords and phrases were first collected. The grammars of language and concern recognition were later significantly improved by evaluating recorded customer conversations. The bot is ideally based on the language of the customers, while internal company technical terms are avoided. By individually designing the dialog system, the voice portal was tailored to the requirements of Signal Iduna. Finally, the functionality of the processes, the comprehensibility of the questions, etc. were tested and further optimized in extensive customer testing.
“More time for excellent service”
Since June 2019, all calls received by the insurer's customer service have been routed through the voice portal. Every day, the portal processes over 5,000 calls. The majority of all callers are immediately assigned to the correct specialist department. The system recognizes customer concerns and, if necessary, performs identification and authentication. Service employees are thus relieved of monotonous tasks and can focus on consultation-intensive discussions. “Thanks to the relief provided by the bot, our employees now have more time to provide excellent service”, says Andrea Schmidt. Customers, in turn, are happy with reduced waiting times and the direct connection to a suitable contact person. Hardly a caller reacts annoyed to having to talk to a bot first.
The project was also so successful because the responsible team was in perfect harmony. A fruitful transfer of knowledge developed between colleagues from IP Dynamics and Signal Iduna during the project period. “The team is an excellent match not only professionally but also personally”, summarizes Schmidt. “Thanks to the seamless collaboration with IP Dynamics, we were able to overcome all challenges and implement the project within the allotted time. ”
The detailed reference report on the integration of a modern language portal at Signal Iduna can be found here.
If you are also interested in automating your customer dialogue, please feel free to contact us.

Dr. Moritz Liebeknecht
IP Dynamics GmbH
Billstraße 103
D-20539 Hamburg


