At OP Financial Group, online chat is quite popular with our customers. We realized this when we first deployed an AI conversational assistant called Viljo for insurance customer relations. Viljo’s launch results exceeded expectations.

Building on this success, we’ve introduced Opotti, a virtual agent for our banking customers. Powered by IBM watsonx Assistant natural language processing technology from the cloud, it understands questions posed in everyday Finnish and pull answers from a knowledge base of banking support information.

Usage has skyrocketed, and today OP virtual agents have already received more than 350,000 inquiries. Opotti can answer nine out of ten questions and automates approximately seven out of ten questions asked without any human intervention.

Learn more about IBM banking solutions

Customers seem to value quick responsiveness when they are in the need of assistance. The virtual agents respond almost instantly even when the contact volumes have been higher. But answering fast is not enough. The virtual agent must prove itself and give correct and helpful answers. We are most proud that our latest poll showed more than 75 percent positive customer feedback for Opotti.

Data science improves customer service

We developed Opotti in a collaboration between IBM Services and OP’s data science unit, one of the largest in Finland. From the start, we were impressed by the virtual agent’s scalability. Opotti answers instantly despite relying on a knowledge base ten times larger than the insurance virtual agent. And like Viljo, it frees live agents to focus on more complex issues that require the human touch.

Beyond answering customer questions, Opotti’s integration with our banking information systems adds a transformative capability: the ability to trigger and orchestrate personalized queries prompted by customer requests. Available to registered customers, Opotti can automatically pull up, for example, customer payment information from within its digital interface. In development are a range of additional services through the chat interface.

The two virtual agents save OP millions of euros by increasing operational efficiency. Perhaps most important, their AI-powered capabilities are transforming the experience of digital customers.

Lessons learned about AI development

The development process taught us key lessens about AI. Although we’re applying AI across the enterprise, the customer relations use case has special merits. Efficiency gains from work-hours reduced and salaries saved are concrete, leading to strong ROI. And, of course, the better customer experience is a competitive differentiator.

Another lesson is that AI development is a collaborative effort among data scientists, software engineers, business experts and technology partners. Top efficiency requires each to perform its proper role. As an example, data scientists do their best work behind the APIs, business experts and AI trainers out front with the user interface.

Senior customer relations agents trained Viljo and Opotti by rating their proposed answers to a range of typical questions. It took just ten weeks for Watson to become capable. And, as more and more customers interact with Viljo and Opotti, our results continually improve.

The right partner for AI transformation

The right technology partner also makes a difference. The IBM team delivered every iteration on schedule and on budget, and the business results exceeded our expectations. We at OP Financial Group look forward to continuing this work as we extend AI transformation to other aspects of insurance, banking, wealth management and business operations.

Watch how OP Financial Group creates a next generation virtual agent named ‘Opotti’ with the help of IBM: