March 1, 2018

Simply speak instead of click—even in industry?

					Digital language assistants from amazon

Alexa for business is to be used in the office environment, for example.
Source: Amazon

Alexa, Siri and Co. help us with small tasks in everyday life. Will digital language assistants soon find their way into factories too?

More than half of the people in Germany have used a digital voice assistant via smartphone, tablet or computer, and another 19 percent can at least imagine its use. This is the result of a representative study by the German Digital Industry Association (BVDW) in November. In the next step, Amazon's Alexa, Apple's Siri and Microsoft’s Cortana want to conquer the business world. Amazon presented Alexa for business at the end of 2017. All possible areas of application cited were in the office area: booking meetings and conference rooms, taking part in telephone conferences, informing IT about a malfunctioning printer in the office, smart home functions (switching off lights, regulating heating, etc.), and ordering office supplies (of course from Amazon).

Language instead of complicated commands

According to Craig Roth, analyst at the IT consulting company Gartner, the announcement of Alexa for business is part of the wider trend to natural language interfaces (NLI) where the end users communicate in their own language rather than having to learn system commands or menus. He is convinced that “a lot innovation could be generated thanks to the development of productivity systems based on NLI.” If this is combined with emerging services such as Polly and Lex, this could be a real opportunity to change the way office and administrative staff to do their jobs. Amazon Lex is a service for the creation of conversational interfaces for voice and text in any application. Polly is the service that converts text to speech with 47 lifelike computer voices in 20 languages.

However, anyone who has ever used or Alexa, Siri or Cortana also knows that the systems do not always properly understand the voice input. When the Zurich startup company Bring linked its eponymous shopping list app with Alexa, for example, it discovered that Alexa interpreted “apricots in a glass” as “alcohol-free grass”. The developers found it amusing. Whoever in an office accesses the wrong phone conference is more likely to be annoyed. In a factory, however, this could have devastating effects such as accidents or unplanned downtime.

The possibilities of language assistants are very limited

However, Alexa for business also makes it clear: the possibilities of language assistants are still very limited. Questions and commands from people are (still) solidly linked with certain software, apps and functions. Consequently, the developers from Bring had to come up with concrete scenarios about what questions end users could ask and what Alexa could answer. This is still far from machine learning and artificial intelligence.

As a result, Siri, Alexa and Co. are only also suitable for factories to a very limited extent in the opinion of experts. This is a chance for European manufacturers. According to Voice Inter Connect, a major challenge remains in distinguishing between voice controls and surrounding noise. The chip manufacturer Infineon stated the currently used language controllers often can only distinguish between a speaking person in the room and an artificial sound source with difficulty. They often consider the loudest sound source to be the decisive one. A new solution from Infineon and XMOS bypasses this problem with intelligent microphones for presence and gesture recognition. A combination of radar and silicon microphone sensors detects the position of the speaker as well as his distance to microphones.

Help without interrupting work

The Viennese company Tablet Solutions has also developed a digital language assistant specifically for industrial enterprises for its Field Management Software WorkHeld. For example, if there are problems with a spindle on a machine tool, the service technician can just ask the tablet what solutions might be possible. The system then checks what faults there might have been—and answers. According to Tablet Solutions, the big advantage of the language assistant is the fact that the technician receives information without having to interrupt his work—and this even without long training periods. The great advantage of language is that it is closest to us, much closer than data glasses, for example.