The ZEW Contact Center 5.0, or CC5.0 for short, is a cloud-based solution developed by SUSI&James that alleviates the workload for companies that make frequent, recurring outbound calls. For example, appointment reminders, scheduling appointments, surveys, and similar tasks can be fully automated. With CC5.0, the voice assistant SUSI—who already serves as a digital employee across various industries—calls customers, patients, and business partners, thereby automating this process that is currently performed by humans. This means that, depending on demand, any number of calls can be routed simultaneously to the appropriate contacts. The advantage of this is that it saves both time and employee capacity. SUSI can learn individual dialogues and thus be flexibly deployed in various applications. To ensure that this application of SUSI functions particularly elegantly, robustly, and efficiently, the CC5.0 is based on the “EVA” AI platform from SUSI&James.
EVA is the first platform in the world dedicated to the development, training, and operation of digital employees. These digital employees are designed to interact logically with humans and communicate effectively. Through EVA, CC5.0 has access to various leading and cutting-edge technologies, such as a dialogue system, a machine learning framework, speech-to-text and text-to-speech components, as well as integration with a cloud communications provider to incorporate the voice assistant SUSI into the telephony system. Artificial intelligence (AI)-based digital assistants speak to respondents in natural language. CC5.0 enables the import of contact phone numbers and allows the voice assistant SUSI to systematically call these numbers using a pre-defined dialogue. This technology has the potential to relieve a large number of people and businesses by having SUSI automatically handle recurring outbound calls. Specifically, this proves very useful in the context of a telephone survey involving a large number of people.
The Research Collaboration
In this proof-of-concept project, we wanted to determine, through a field experiment conducted in collaboration with the Centre for European Economic Research (ZEW), whether voice assistants are suitable for use in opinion research, specifically for automated telephone surveys. The aim was to ascertain how the respondents' acceptance of the voice assistant compared to human interviewers. Furthermore, the project's objective was to use our SUSI system to automatically call numerous young companies in Germany and conduct a short survey on the topic of COVID-19. Crucially, this involved understanding how the pandemic has impacted the operations of German businesses. The implementation and analysis of this scenario took place over the past few weeks. This initial scenario served as an acceptance test for artificial intelligence and represented a breakthrough for the use of AI as a human-to-human interface. The CC5.0 product enabled ZEW to make multiple calls simultaneously and complete the entire survey within a short timeframe, instead of undertaking a large-scale project with a call center. For this to work, ZEW provided CC5.0 with the corresponding telephone numbers to be contacted. Furthermore, a pre-defined dialogue should already contain information that SUSI should request. In this regard, the ZEW had 500 registered companies with complete data including telephone number, managing director, and location.
With a single click, the CC5.0 automatically called all defined contacts, and the digital assistant SUSI independently conducted the survey using a sophisticated dialogue. First, she introduced herself and explained to the respondent what the survey was about and what role they played in it. She then continued with the pre-written dialogue, asking three simple questions that could be answered with “yes” or “no.” Specifically, the following questions were asked:
“Has the COVID-19 pandemic had a predominantly positive economic impact on your company so far?”
“Have you received or applied for assistance or support from the federal government, state governments, or local authorities for your business?”
“Are you considering changes to your strategic direction for the post-pandemic era, or are you already working on them?”
SUSI sammelte alle benötigten Daten im Gespräch und speicherte diese anonymisiert in einer sicheren Datenbank. Nachdem alle Umfragen abgeschlossen wurden, konnten die Ergebnisse eingesehen, ausgewertet und weiterverwendet werden. Hierbei wurde deutlich, dass ein solcher Prozess, wenn er von Menschen ausgeführt würde, eine sehr lange Zeit in Anspruch nähme. Aus diesem Grund bat das ZEW-Outbound-Projekt einen enormen Mehrwert und kann die Beziehung zwischen künstlicher Intelligenz und Menschen künftig verändern.
