Artificial intelligence has been in the spotlight for several years, and the topic has truly made a breakthrough with the arrival of ChatGPT. Although new applications have helped the development of AI to take major leaps, artificial intelligence itself is nothing new to the debt collection business.
Quite the contrary: Lowell has developed and tested various types of models for machine learning and artificial intelligence for nearly two decades. Currently, AI-based models are widely in use throughout the entire debt collection process.
– AI does not have any intrinsic value for Lowell as such, unlike efficient business operations and providing outstanding service, for which AI is an excellent tool, says Olli Helttula, head of digitalisation for Lowell in the Nordics.
AI boosts backend systems
In the debt collection business, artificial intelligence operates mainly in the background; it increases the efficiency of backend systems and raises the level of automation. Efficiency may sometimes be considered a negative attribute, but Helttula notes that it is usually visible also as higher quality.
– Automation is reliable because all operations are performed in the same way, without any deviations caused by humans.
A simple example of enhancing operations is the recognition of texts and images. When, for example, people are no longer needed to read the content, the processing of printed letters becomes faster. Depending on the number of letters, even up to thousands of working hours can be saved.
If a statistical model helps to make, for example, five per cent better decisions, it can mean savings of millions of euros yearly.
Artificial intelligence is also used to classify masses of data and search for anomalies. It can help to develop the logic of decision-making and create statistical models to help in making decisions. The impact is significant on the business of a company the size of Lowell.
– If a statistical model helps to make, for example, five per cent better decisions, it can mean savings of millions of euros yearly, Helttula adds.
Lowell has been developing and testing artificial intelligence for nearly two decades. AI-based models already drive the entire collection process.
Virtual assistant improves the service experience
As a promising trend in artificial intelligence, Helttula highlights virtual assistants that help Lowell’s employees in their tasks. A virtual assistant can, for example, provide support during a customer service call.
– At the same time as the caller explains their situation, speech recognition can search for necessary information on the customer without needing customer support to do it themselves.
OpenAI’s ChatGPT has revved up the debate on artificial intelligence in the debt collection sector.
– ChatGPT is a huge step forward for artificial intelligence. Especially AI-based applications demonstrate how groundbreaking the technology is.
According to Helttula, AI is currently suited for answering simple questions or assisting in searches but is still not reliable enough for handling more demanding discussions.
be used to chat directly with customers, as the generated information is not reliable enough. Also in the future, Helttula believes that AI will continue to assist customer support, rather than replace it.
True artificial intelligence is yet to arrive
Such artificial intelligence that can handle everything without human supervision will still take some time to emerge, Helttula says. He points out that we should be cautious about applying AI to all kinds of tasks.
– Companies may perceive chatbots and other customer service applications as things that automatically enhance operations. But when used too extensively or incorrectly, they only worsen the experience of the end customer.
Companies may perceive chatbots and other customer service applications as things that automatically enhance operations. But when used too extensively or incorrectly, they only worsen the experience of the end customer.
Chatbots that utilise AI are suitable for sorting out simple matters, but more challenging cases remain the responsibility of human beings. So far, chatbots are not sufficient and often understand things incorrectly in conversations.
If a customer needs to explain their case to a robot that does not understand, they become frustrated. In Helttula’s view, this is not a responsible way of utilising AI.
He reminds us of the fact that artificial intelligence is not applied to the debt collection business because of the hype surrounding AI, but because it is genuinely useful and helps to do things better. He believes that artificial intelligence is here to stay in the business.