The well-being and health of employees in the workplace are relevant aspects for the economic and sustainable performance of a company. If these are disturbed, this can lead to considerable and occasionally “irreparable” effects; For example, a high employee turnover, burnout incidents and the loss of a company’s profitability. Causes of such a disturbance of well-being in the company is varied and requires a detailed analysis.
“Information Overload” through systems and technology
For example, the high number of work tasks or tasks to be performed in the professional environment and the associated time pressure overwhelm a large number of employees and executives. Among other things, the time pressure arises as a result of overlapping information and tasks, which means that many employees find it difficult to meet the quality requirements of day-to-day project and administrative tasks. The predominantly understaffing of companies also leads to an increase in the burden. The high number of different software solutions also overwhelm employees who have to remember numerous usernames, passwords and software processes in order to reach their “goal”.
Digital assistants as information filters
To solve the problem of the flood of information and tasks as well as the large number of software systems, work routines would have to be simplified. In plain terms, this means either using fewer tools and systems, or using personal assistants that act as filters to reduce complexity. Both are not easily possible in the classical understanding. The key to problem solving is the use of new technologies that are switched between employees and all of their tools and software solutions to reduce complexity right away: digital assistants, as a conglomeration of disparate technologies, can help.
Technological developments as the basis of digital assistants
The endeavor – to make the work of the employee more enjoyable with the help of technology – is very complex from a technical point of view: on the one hand, one has to understand the challenges in the professional environment and the solution “digital assistant” has to fit into the company as well as the software solutions used. On the other hand, a user-friendly and multifunctional user interface as well as a linkage of devices and smartphones is required. In the best case, the digital assistant should be able to (semi-) automate non-value-added and recurring tasks. These challenges are solved by the digital assistant “Neo” with different technologies and approaches, which are explained below.
Mobile computing as a key technology
In mobile computing, the software and devices used are designed for mobile use, including notebooks, tablets, smartphones and smartwatches and their corresponding software solutions. Consequently, not only the hardware is understood by mobile computing, but also the software used. With the help of mobile computing, users can move around flexibly and without giving up their software solutions. Above all, mobile computing is supported by numerous apps that enable simple and fast operation via mobile devices.
Mobile computing has also become established in the professional environment. The term “Enterprise Mobile Computing” uses similar technologies and approaches as those used privately in professional environments. Nevertheless, there are differences to private use. In the professional environment or enterprise sector, mobile computing is used to increase economic value added and must be integrated into the complex corporate world. Therefore, the requirements for mobile computing in the professional use differ in the private environment in part. For professional use, the mobilization of existing digital business processes and the development of new business areas through the use of mobile computing as part of digital transformation are two relevant scenarios.
Artificial intelligence as a basis for digital assistance systems
In general, the field of “Artificial Intelligence” (AI) is a branch of computer science in which with the help of IT, an “intelligent behavior” of the software, devices or machines can be achieved. Behind the term of the AI is a large number of diverse methods, procedures and technologies. Every approach and every method brings with it a different variety of possibilities, so that a different process is selected depending on the application scenario in the company. The use of Artificial Intelligence always requires a data model, which is the basis of “intelligence”.
The breakthrough: data+computing-power
The field of Artificial Intelligence has been researched for decades, but now the breakthrough seems possible. Among other things, this is due to the fact that there is a high volume of data in companies and this data volume can now be stored. In addition, a higher computing power has been possible for a long time, enabling fast calculations (almost in real time) and as many operations as possible in a very short time. With data generation and storage and high-performance computing, companies can use a variety of Artificial Intelligence techniques to calculate more complex situations and configurations.
Strong and weak AI
A “strong AI” is understood to mean a theoretical Artificial Intelligence whose goal is to imitate human intelligence but is currently outside the technical possibilities. Behind the “weak AI” construct, procedures are aggregated that can make intelligent decisions in certain areas: for the (semi) automation of processes and tasks.
Conversational Interface: A conversation with the computer
A Conversational Interface refers to the user interface and operating concept of a software that takes the form of a dialogue with the computer program. In contrast to traditional, graphical user interfaces, the extensive interaction with peripheral devices such as the mouse is dispensed with here. Conversational Interfaces thus reflect the current trend from UI to UX design, whereby the content not only determines the visual design, but also becomes the medium itself.
By largely eliminating visual elements, Conversational Interfaces can be displayed correctly across devices and platforms on desktops, smartphones and smartwatches – and thus be developed faster and cheaper. The underlying interaction design can even be adapted to devices without a screen (such as speakers with voice control). By focusing on the content as a visual medium, it can be adapted to the user’s needs and enriched with contextual information. The user can always be supplied with the information that is relevant to him at that moment, without the need for additional operating elements.
Rapid Prototyping with Conversational Interfaces
This makes Conversational Interfaces suitable for rapid prototyping and simplifies user behaviour analysis through real-time feedback. New functionalities do not have to be provided extensively via updates and installation processes, but can always be supplemented and improved on the server side. Conversational Interfaces put the user in the focus of software development and are ideally suited for the development of the digital assistant Neo.
Speech recognition and output
In humans, spoken language is the most natural form of communication. The “speaker” selectively selects a certain sequence of words to convey a message.
The Speech Recognition (Speech-to-Text, STT) has the goal to reconstruct the uttered word sequence error-free. To achieve this, identify the basic tone units in the word order. Using a list of possible sound units, a complex language model is created that recognizes and evaluates certain structures based on our existing language knowledge and correctly identifies the spoken words in the word order.
The Speech Output is called “Speech-Synthesis” (TTS), which is the artificial production of human speech. In doing so, a Speech Output is generated from an existing text. Two methods can be used for this: The output of speech through so-called “samples” (voice recordings) and the so-called physiological (articulatory) language modeling, which is generated by high computing operations by servers.
Error rate at almost human level
Until recently, Speech Recognition and Speech Output was still a rarity, but that changed within a very short time. Among other things, this is due to the fact that the acceptance of Speech Recognition and Speech Output has increased because, above all, the error rate in Speech Recognition has been significantly reduced. In recent years, the error rate has fallen from more than 20% to less than 5%, with a downward trend. Also indistinct expressions, ambient noise and even dialects are almost unproblematic. Through this development, Speech Recognition and Speech Output will radically change the conventional interaction with tools, software and entire system solutions.
Enterprise Application Integration 4.0
The concept of Enterprise Application Integration (EAI) has been around for some time and was touted as “the” solution for enterprise-wide integration of all business functions along the entire value chain of a company, especially at the beginning of the millennium. EAI aims to connect all applications installed on disparate-distributed platforms to one another through data and business process integration, resulting in fully integrated business execution. This allows data from different sources to be merged. However, this requires clearly defined interfaces in order to allow complete data transfer.
The trend towards cloud solutions and the consistent provision of APIs (Application Programming Interface) brings a renaissance to the EAI concept. Systems and tools can be connected more easily and quickly, linking entire business functions along the entire value chain.
On the shoulders of the giants
Neo as a result of influencing factors and key technologies
The above-described influencing factors and key technologies created the digital assistant for Business, Neo. Neo can be used on smartphones (iOS and Android), on desktop devices (macOS, Windows and Linux), on data glasses (Android) and on Pepper-Robot and thus enables the user or employee maximum flexibility and mobility.
With the help of Artificial Intelligence (Natural-Language-Understanding) and a Conversational Interface (by text or speech based on a conversation engine) Neo understands and processes the inquiries and interactions with the user and supports the employee context-based in his daily work routine. Instead of clicking in various systems and tools in the menu navigation, the employee asks a question and Neo replies using Speech Recognition and Speech Output.
The technical actions to be carried out are coordinated by means of a central switching and integration platform. Neo can connect third-party systems used in 18 different programming languages in the company, thus enabling maximum flexibility in the integration of third-party systems. In the long term, Neo enables (semi-) automation of various manual processes and, as a central orchestrator, provides a knowledge network in the company.