More comfortable working through speech with the use of Artificial Intelligence (AI) in industry

We live in a time when we already use Artificial Intelligence, abbreviated as AI in the following, in many areas. The average consumer often doesn’t seem to notice. How could they?! AI is not tangible and you cannot see it with your eye. The use of AI helps us, for example, with research on the Internet, simplifies and accelerates business processes or even takes over tasks in the household.

In private areas, voice assistants are being used more and more, such as smart home systems or voice assistants like Siri and Alexa.

With simple questions like “Hey Alexa, what’s the weather going to be like today?”, “Hey Siri, when is the next train to Berlin?” we can get answers in seconds. Even the front door can be opened from a distance. Not to mention self-driving cars.

All you need is an intelligent assistant app on your smartphone, laptop or tablet.

Keeping in mind that there are so many people using voice assistants, we have set our focus on intelligent voice assistants for businesses of all sizes and types, providing access to modern AI for more convenience in work processes.

In this report, we would like to specifically draw attention to our self-developed inspection assistant. This assistant is used in industry, inspection halls or automotive workshops. It speeds up and simplifies tedious and time-consuming work and interruptions are reduced, as work here is “hands-free”.

Let’s start with the problems that often happen in industry:

During an inspection (e.g. the main inspection of a vehicle), the vehicles are examined by inspectors according to a predefined inspection plan. The defects found are recorded in a structured manner on a tablet or laptop. The entry is largely done with the hands. In the worst case, protocols are still noted down with pen and paper.

This is where we immediately see the problem or challenge for the employee. Here, for the most part, error-prone, and time-consuming paperwork associated with a lot of “interruptions” must be completed.

Our solution:

Through a voice control, the recording can be done much more comfortable, faster and error-free.

By integrating the inspection assistant app on the website, in the phone or in the company’s own app, the entry of defects can be made, for example, via a headset by voice input while the inspection is being carried out. This greatly simplifies the work process and saves a lot of time and money in documentation.

How does it work?

KIDOU, is a language kit we developed that includes many AI agents. We have adapted and developed the AI language model at an industrial company until it is ready to go live.

The AI agents are scalable, extensible, and adaptable, and therefore can be used in many other industries, such as medical, nursing, sales, and more.

KIDOU can transcribe the spoken word in seconds and transfer it to the customer’s backend system, so no further manual steps are required. The data is available for immediate use. During voice input, KIDOU provides active feedback in the event of an error, for example, if the input is incomplete.

It is also possible to control a customer app individually by voice, e.g. navigation through the menu, it is also possible to take a photo or enter free text information.

What about special technical terms, which are very common in the industry?

For a voice input of inspection results, technical terms are essential in this application and some language systems do not always understand these technical words. To ensure that KIDOU can reliably recognize the necessary terms, the language model is adapted, optimized and trained in a training process.

The integration and implementation process is carried out with the customer. The necessary training data is both automatically generated by us and collected during operation (if the customer agrees). During the training process and afterwards, our employees work hand in hand with the customer.

How does the assistant handle noise and background sounds?

One challenge is that there is usually a lot of noise in industry or in workshops. With the latest voice technologies and intensive training, voice input can be recognized reliably despite disturbing noise and background sounds. 

Conclusion

Voice is the fastest and most natural form of communication. The inspector can enter his defects and reports handsfree simply by voice while performing the inspection on the vehicle or machine. Thus, he can accelerate and simplify his work processes and high time and cost efforts are reduced.

Since the app runs on all systems, whether iOS or Android, it is quickly adaptable to the domain, individual and adaptable. For companies with high data protection requirements, the data can be processed in the in-house server (on-premise). In addition, AI solutions are also available in the cloud, hybrid or on-edge.

On our YouTube channel, we have tested our inspection assistant (prototype) in a car workshop.