Artificial intelligence enablers: the building blocks of the future
AI enablers help companies to combine services in a rather easy way to solve their individual business challenges.
Tim Cook, CEO of Apple, recently stated that Artificial Intelligence (AI) is “horizontal in nature, running across all products” and is used “in ways that most people don’t even think about.” At Swisscom, we strongly agree with Tim Cook. We believe that AI will be integrated into every application in the future. Without AI, an application will appear stupid and therefore be worthless to the user.
The various benefits of AI enablers
Many companies do not (yet) have AI skills and technology in place to make their applications smarter. To help them benefit from AI, we packaged our AI capabilities into several so called AI enablers. These are standalone micro services, running on secure infrastructure and accessible through APIs. We think of our AI enablers as the building blocks of the future: Our enterprise customers can combine them in many ways to solve their business challenges.
Furthermore, we use AI enablers within Swisscom to make internal applications and processes smarter and more efficient (for more detail, have a look at the ICT Blog post “Artificial intelligence rends customers happy“).
At Swisscom, we currently use and offer the following AI enablers:
- Entity recognition
- Topic recognition
- Sentiment analysis
- Trend Detection
- Keyword Analysis
- Model Training
More than just one use case
An AI enabler can be used in several use cases. Two examples: The same sentiment analysis enabler might be used for identifying the feelings of a customer giving feedback on a product and of an employee giving feedback in a satisfaction survey. The topics recognition enablers can be used to cluster thousands of customer feedbacks according to topics. They are also able to identify the most important topics and ideas addressed in the yearly employee satisfaction survey.
Currently, several competitors are flooding the market with AI offerings. In order to provide added value to our customers, our AI enablers are based on three principles:
- Cutting-edge and open source algorithms: We are currently witnessing an extremely fast evolution of AI algorithms. We employ and modify cutting-edge algorithms from top universities and companies in a way that our clients are able to use them fast and easily. Furthermore, we choose an architecture for our AI enablers that allows us to quickly integrate new algorithms. This approach allows us to be at the forefront of AI and allocate our resources to business activities. That is how we can provide our customer with more value than our competitors, such as use case specific model training.
- Support of local languages and industries: A majority of the cutting-edge algorithms are trained to work solely in English, Chinese and Spanish. However, most people in Switzerland speak German, French or Italian. Therefore, we provide our customers with pre-trained models for these languages. We also take into account that the Swiss industry is historically strong in a few key industries and are equipped to provide pre-trained models for them.
- Reliable infrastructure: While we embrace the numerous startups in the field of AI, discussion with customers showed their concern about startups’ troubles to provide enterprise-grade, reliable, and secure infrastructure. That’s where we bring the full force of Swisscom data centers into play. Furthermore, by making sure that data stays within the perimeters of Switzerland, we follow several companies’ governance regarding data security.
Our AI enablers allow us to tailor an AI solution for our clients’ needs. Nevertheless, our set of enablers is not yet complete. Based on the needs of our customers we are currently working additional enablers. For example, we plan to release the speech to text, machine translation and generic chat bot enablers next year.