Faction XYZ: Towards Enhancement of Businesses with Machine Learning

Jos Polfliet, Head of Applied Artificial Intelligence

"Builders.Not Talkers." That is the tagline for Faction XYZ – an applied AI engineering service provider.The firm is in the business of empowering organisations with a vision and budget to use A.I. driven automation abilities for an economy where applications of machine learning and data science will determine the competitiveness of their clients’ business. Based in Antwerp, Belgium, Faction XYZ constructs machine and deep learning algorithms in the area of natural language, computer vision and sensor data for the telco, retail, financial and automotive industries. “We focus on delivering value by applying machine and deep learning to enhance productivity, streamline operations and data flows of an organisation, or come up with competitive products”, states Niels Van Weereld, Head of Business Development of the firm.

Faction XYZ builds intelligent and smart digital assistance that help manage conversations, automate email conversations and, in some cases, automate voice conversations over the phone where digital agents listen and transcribe. These processes have many benefits as there are no long waits in queues and help is available immediately, round the clock. Faction XYZ also focuses on sensor data – in a manufacturing context. Cases include developing predictive maintenance framework for automotive companies where clients shift to ‘car as a service’ model. For instance, clients will get to know when a car battery is about to die and when a tire will wear out. This ‘car as a service’ model ensures things are replaced much in advance, saving people from an messy situation later.

Besides, the firm has developed chatlayer.ai, one of Europe’s fastest growing enterprise-graded conversational management system that enables the business-side of large organizations design both sequential as well as natural language driven flows effortlessly. There are few important value drivers, automation personalisation, and optimisation. Initially, a target is defined, and conversations are optimised accordingly. The platform is entirely extensible using off-the-shelf plugins that make it apt for any project. “We do not compete on cost; we compete on quality and believe in delivering the best solution for our clients. Even if this means we have to turn them down if we feel machine or deep learning bears no added value for them. We are a highly pragmatic and client-centric company, which is the reason for our success so far”, asserts Van Weereld.

Niels Van Weereld, Head of Business Development

Outcompeting incumbents in quality and accuracy with their unique natural language models, such as Dutch and French, one of their many clients to benefit from the solutions provided is a large telco in Belgium. Faction XYZ came up with two things: The first was a sales navigation chatbot. The client’s website has a chatbot that acts as a sales assistant to help customers navigate through their product portfolio. Since their product portfolio is not very straightforward, the chatbot guides and provides customers with vital information to choose the right product. This process increases conversation rates by large amounts compared to just having people browse through the website. Second, the firm has helped the company eliminate some of their troubleshooting, such as their internet connection issues. “When a company approaches us, we try to understand their problems first, and then fit in machine learning, data science and artificial intelligence to solve their issues,” says Jos Polfliet, Head of Applied Artificial Intelligence.

Factions XYZ is expanding geographically and providing services to customers in Western Europe – an important market for them– as well as in the rest of Europe and North America. As for their product Chatlayer.ai, the firm has a solid roadmap in place that continues to launch improvements and new features for 2018 and beyond. “With enough useful data to build models and gain insights and with improvements in algorithms, this is the best time to utilize deep learning,” concludes Van Weereld.