IPD ×
Signal Iduna

A new colleague within customer service

With the help of artificial intelligence, a significant part of Signal Iduna's customer dialog was automated.

Industry
Insurance
Employees

> 10.000

Turnover

approx. 6.5 billion euros

Thanks to the smooth cooperation with IP Dynamics, we were able to overcome all challenges and complete the project within the scheduled time.

Andrea Schmidt

Project Manager at SIGNAL IDUNA

The Challenge

A question of routing

The Signal Iduna Group was formed in 1999 following the merger of the former Signal Insurance Group of Dortmund and the former Iduna Nova Group of Hamburg. With over 10,300 employees across Europe, the insurance Services company is committed to addressing the growing needs of its customers and communicating with them as equals. So it comes as no surprise that first-rate customer dialogue is paramount for Signal Iduna. A key focus here is its telephone service, which often represents a direct point of personal contact between the Company and its customers. The challenge consists in connecting each and every individual caller to the most suitable contact person for the matter in hand. A company the size of Signal Iduna has many different lines of business and a vast array of telephone numbers. Customers often dial the first number they come across. In such cases, an intermediary routing facility can recognize the reason for a customer’s call and quickly help connect him or her directly to a suitable contact person.

Back in 2016, IP Dynamics installed a Voxtron Communication Center (VCC) for Signal Iduna on an existing Skype for Business infrastructure. At the end of 2017, the insurance company then decided to integrate a modern voice portal into the system. Pleased with the positive experience with Voxtron, it commissioned IP Dynamics to implement the new project. At the time, Signal Iduna was using a simple Interactive Voice Response (IVR) system, where callers could press buttons (DTMF) to navigate through a two-stage voice menu. However, select buttons are limited as they offer too few selection options. Also, callers had to make what was for them the difficult decision about the line of business to which their issue related. As well as IVR, Signal Iduna also had a conventional telephone switchboard through which employees received around 650 additional calls that were not routed via IVR.

During the concept phase, IP Dynamics’ vast experience from similar projects paid off. IP Dynamics gave us valuable tips on fine-tuning voice recognition and dialogues. But also on matters that we hadn’t really considered before.

Andrea Schmidt

Project Manager at SIGNAL IDUNA

The Solution

All routes lead via the voice portal

The new voice portal project aimed to reduce customer-call forwarding rates and shorten caller waiting times. “If calls go to the wrong contact person, meaning issues have to be re-explained multiple times, customers get annoyed and service employees waste time”, explains Andrea Schmidt, Signal Iduna’s manager for the project. “We wanted to introduce optimized routing to eliminate this situation, which was unsatisfactory for everyone.”

The existing system of IVR navigation by key selection thus had to be replaced by one based on natural voice recognition. The new voice portal needed to greet callers in a friendly manner, recognize the issue they wanted to discuss and – if necessary – perform authentication by asking for their insurance policy number and customer data. The extensive data protection check, which previously had to be carried out by an employee, was to be automated through the new voice portal. Following successful routing, a CRM interface was to provide the relevant service employee at a glance with all information given to the voicebot by the customer. In addition, a “smartlink” functionality was to be integrated so that the voicebot could where appropriate refer customers to self-service features on the company’s website. However, the system was not under any circumstances to be used to circumvent dialogue with employees. Where an enquiry could not be successfully dealt with by the voice portal or a caller declined a self-service option, the voicebot was, where possible, to connect the caller directly with a human colleague.

“The particular challenge we faced with this project was thinking through all possible dialogue paths in advance”, says Dr Jürgen Haas, IP Dynamics’ manager of the project. “This required us to map the entire spectrum of phrases that customers use to describe their various issues. To guarantee successful customer dialogue, a voice recognition System must ultimately be prepared for all eventualities.”

The voice portal project was launched in January 2018. During the first six months, the project team devised an extensive concept. Staff in the various specialist Departments at Signal Iduna engaged in constant dialogue, managing technical insurance-related and process-based inputs and specifying requirements, while the experts at IP Dynamics supplied technical advice. During this process the formal basis was clarified, Works Council approval was obtained, data protection issues were resolved, and the routing Targets for the new voice portal were defined. “During the concept phase, IP Dynamics’ vast experience from similar projects paid off”, explains Andrea Schmidt. “IP Dynamics gave us valuable tips on fine-tuning voice recognition and dialogues. But also on matters that we hadn’t really considered before – such as the time limit allowed for callers to formulate their answers.”

IP Dynamics started technical implementation at the end of 2018. The first task was to create a technical basis using the latest voice and customer interactive technologies to enable natural dialogues. The required technology was integrated using the Dynamic Flow (DFW), IP Dynamics’ central workflow engine with voice portal functionality. As well as enabling connection of a voice recognition engine, the DFW also made it possible to integrate the CRM system used by Signal Iduna.

To overcome the aforementioned challenge of considering all potential dialogue paths and the entire spectrum of callers’ phrases, one of the first tasks was to collect keywords and phrases. However, a practical analysis of specific case studies was even more important than theoretical considerations. Analyzing recorded conversations with customers helped significantly improve the grammars involved in voice recognition and in identifying the reasons for calls. During the first quarter of 2019, extensive customer tests also verified the functionality of the processes, the understandability of questions, etc. This made it possible to optimize certain dialogue-related intricacies. For example, many customers answered “Signal Iduna” when asked the introductory question “What insurance are you phoning about?”. However, the question of course referred to the customer‘s insurance policy, not the company. To Prevent such misunderstandings, questions were made more precise or supplemented with specific examples. In this case, for example, the question was supplemented with “For example, say ʻhealth insuranceʼ”. Ideally, the bot is tailored to customers’ language, while avoiding internal company expressions or specialist terms. Individual dialogue structuring  nsured the voice portal was fully tailored to Signal Iduna’s requirements. In addition, a maximum of two repeated questions was  et in order to make dialogues as smooth as possible for callers. If it has not been possible to recognize the issue at hand during the course of this process, the call is routed directly to an employee. In principle, all dialogues are designed to ensure there is always the Option of speaking to an employee.

An initial service telephone number finally went live on the new voice portal in March 2019. Quality control was expedited in parallel to the subsequent rollout, and the System was further optimized. The rollout was completed at the end of the second quarter.

Today, Signal Iduna benefits from a wide range of advantages:
Automatic request recognition, identification and authentication
Shorter waiting times for customers
Relief for employees
The Result

“More time for excellent service”

Since June 2019, all calls received by the insurer’s customer service department have been routed by the voice portal. The portal processes over 5,000 calls every day. Most callers are immediately routed to the correct specialist department. The system recognizes the reasons for customers’ calls and performs identification and authentication where required. Relieved of such monotonous tasks, service employees can now focus on support-related calls. Man and machine work hand in hand, so to speak. By ensuring a transparent structure for the voice portal from the outset and providing training to prepare employees for the new technology, Signal Iduna was quickly able to dispel any isolated misgivings. “The reduction in workload provided by the bot has freed up more time for our employees to provide first-rate service”, says a delighted Andrea Schmidt.

Customers themselves are pleased with the reduced waiting times and the fact that they can be directly transferred to an appropriate contact person. Scarcely any callers have expressed annoyance at first having to speak to a bot. In future, peaks in demand can be cushioned by the “smartlink” functionality. The self-service feature offers customers fast initial assistance even when call volumes are high.

The project’s resounding success was also down to the well-matched team. Since the beginning of 2019, colleagues from IP Dynamics and Signal Iduna have worked together at the insurance company’s head office in Hamburg for three days a week. This has led to a productive Transfer of knowledge over the course of the project. For example, Signal Iduna staff have now learned how to write voice recognition grammars independently. “As well as being a great technical match, the team also gets on well from a personal perspective”, concludes Andrea Schmidt. “The smooth cooperation with IP Dynamics has enabled us to overcome all challenges and complete the project on schedule.” Dr Jürgen Haas is also pleased and is looking to the future: “The voice portal offers huge potential. It has been implemented as a complete system, but can where necessary be further expanded to include additional features and functionalities.” One such possibility is full or partial automation of certain processes through direct phone-based self-service features. Self-learning AI can also be integrated to cover even more applications. And the voice portal can of Course be scaled to requirements and extended to include other areas of the business.

Find out more details about this case in our detailed reference report

Customer support with Microsoft Teams: Native integration of your service center

Telephony, video, chat - transform Microsoft Teams with Dynamic Workload into your service center of the extra class!

Ready to optimize your business?

Would you like to use our solutions in your company? Just contact us. We are looking forward to seeing you!

*Mandatory information

We only use your data to process your request. The transmission is encrypted. In this regard, please also note the data protection information.

Thank you so much! We've received your message and will get back to you shortly.
Whoops, something went wrong. Please try again or contact us by phone.