Complement a manual with a voice operated Q&A system
Starting point - car
A premium car company experienced that the increasing sophistication of the vehicles was accompanied by a continuously expanding size of their manuals. Infotainment systems, navigation systems, climate control systems, etc. required more and more manuals. At the same time, they were acutely aware that consumers rarely ever read manuals – or where able to use all the functions that were included in the cars.
The problem was recently investigeted by Continental: 40% of 2000 car owners claimed they were unable to use all the functions in their car.
Setting up the service
The focus of our collaboration was the car infotainment system. A free-format voice-controlled system was built based on the content of the manual – “if this is the answer – what is then the question?” The system was installed on a remote server. Drivers called a number, asked a question in free format, and had the answer read back.
Result, experience gained
The trial, which was launched in 2004 (seven years before the launch of Siri, ten years before the launch of Alexa), was one of the earliest free-format voice support systems deployed. Users asked questions about their car, in their own language and with their own words, and were provided answers using synthetic speech.
In-car free-format voice-operated media player
Starting point
Our first smartphone application was DialogueTech’s Personal DJ – a voice operated media player in a mobile phone. The application was launched in 2006 - stretching the then-available-technology to the limit.
Setting up the service
At the time when our first “Personal DJ” was developed smartphones were very different from today – obviously processing power and memory was much less potent – but the biggest obstacle was that there was no architecture for adding 3rd party applications. The challenges of finding the music metadata, extract it and put it in a database, and build a voice-operated NLQ-system that could initiate the desired music were great.
Result, experience gained
Already a decade ago Ergo would run with excellent performance on a smartphone. Also – the sourced voice recognition had already then good performance – today the performance would of course be even better.
The attached video illustrates an early version (2006) – the system is not integrated in the car but runs on a local PC operated from a headset. As you can see it can handle a diverse set of questions and generate the desired playlist.