The commercial success of voice-controlled systems like Amazon Alexa and Google Home,

coupled with advances in natural language processing (NLP) using artificial neural networks,

creates new opportunities for business-to-customer interaction.

Automated voice dialog systems can be used at low cost thanks to established standards and

open access APIs of the major platform providers (Microsoft, Amazon, Google, Facebook).

Businesses want to leverage this savings potential and improve their availability around the


Within the scope of the work, it is to be evaluated whether a combination of state-of-the-art APIs

and/or open source software in the area of text-to-speech and speech-to-text can be used as an

commercially viable dialogue system for enterprises. An evaluation framework must be

developed to review the currently available systems.

In addition, the most suitable components will be implemented in a prototype. Together with

business partners, the work can then be used to provide an outlook on possible applications,

such as an agent for automatically scheduled appointments for service-providing businesses.

An existing workable prototype programmed in Python can be used which integrates various API

services and makes them available as a separate cloud system. The system is also connected to

the phone network via an interface, so that an automated dialogue on the telephone

(independent of end devices such as Amazon Echo) can be implemented.

The master thesis is supported by nitro ventures GmbH, a Münster-based technology startup. If

you have any questions, Dr. Ulrich Wolffgang is available via e-mail to