The Language Technology Industry Summit 2017 will illustrate how language technology addresses challenges in banking and insurance. Cognitive computing leader Expert System will provide insights and case studies from the financial sector. We have interviewed Daniele Cordioli, Presales Director at Expert System:

LT-Innovate: Expert System is a European leader in cognitive computing and text analytics, in particular for banking & insurance. Could you briefly describe the origins of your company and the main phases of its development up to now?

Daniele Cordioli: Expert System was founded in Modena, Italy by three university colleagues determined to show the world that intelligent software could be created in Europe. That was still at a time when the convergence of linguistics and technology was subject of study in research and academic circles, far from the market.
There were no start-up incubators or angel investors to spur growth, so we relied on many, many hours of hard work to bootstrap the business with what we had. After licensing our early technology to Microsoft, we were able to fully extend our vision as software that could actually understand the meaning and context of language. The effort produced one of the first semantic analysis platforms and led to our patented Cogito technology.
We entered the Italian stock exchange in 2014, and soon after, we completed the acquisition of two divisions of the Spanish semantic technology company iSOCO and of text analytics leader TEMIS. Then, we launched the international cyber defense electronics company CY4GATE and a new subsidiary in the US to meet the growing needs of our private and public sector clients.
Today, we are a leading cognitive computing and text analytics company that provides an ever-increasing number of applications in multiple languages to some of the largest and most admired companies and government agencies throughout Europe, the Americas and the Middle East, and our prospects are still growing.

LT-Innovate: Banking & insurance is a major consumer of cognitive computing. What is driving growth in this sector today?

Daniele Cordioli: Recent studies seem to point out, and we agree with them based on our own analysis and experience, several key factors that are driving change in this sector today. We would emphasize the following ones.
The first one deals with the increasing availability of channels like online and mobile banking, which have joined more traditional ones like phone calls. The heterogeneous modalities (text, voice, even scanned documents and images) of such channels and the overall volume of data resulting from the consequently higher user interaction bandwidth requires new ways to process all the resulting data in order to assist the customer and produce value.
Another important factor has to do with the way customers consume information nowadays, increasingly looking for digital interactions that are simple yet appealing, highly personalized and context-aware so that the need of the moment is served quickly without cumbersome intervention from the service provider. This leads to cognitive and language-intensive applications that relieve the customer from actually inspecting large document collections and lower the entry barrier to the comprehension of such documents by non-experts.
Finally, the regulatory compliance and the current macroeconomic environment have driven financial stakeholders to seek incremental profitability from product and service innovation deployed via digital channels. Return on equity and assets are close to their historical minima. However, such profitability can only be achieved via efficiencies derived from operational changes and process optimization.

LT-Innovate: What are your customers main LT-related challenges in the field of banking & insurance?

Daniele Cordioli: Mobile and online banking achieve an elevated level of user experience when language is employed as a primary interface. Understanding language as expressed by the customer and acting upon the request achieves a high level of customer satisfaction and loyalty.
The same holds true for service. When customers express their needs within the context of their understanding, in their own words, this offers the institution a unique opportunity to satisfy the need as expressed. No matter which channel the customer is using—self-service systems of live customer support—being able to respond without an elaborate or repetitious translation into the language of the bank or insurance company is perceived by the customer as friendly and empathic.
In banking and insurance - besides improving customer interaction and experience - we see cognitive and language-intensive applications as driving process optimisation, notably in claims management, contract clause comparison, risk screening, compliance and anti-money laundering. Our product, Cogito, is extremely powerful when applied to identifying and exposing abstract relationships, supporting search and correlating content, which is why a growing number of financial institutions rely on our technology.

LT-Innovate: What is the state of competition in this sector? New fintech start-ups versus established tech suppliers?

Daniele Cordioli: If you look at the investment poured into fintech startups in the last few years (approx. $20billion per year since 2015, according to Forbes) it is obvious that technological innovation is one of the keys to remaining leader of the pack in such a competitive sector. This is something that has always been crystal clear for us since the very beginning and has enabled Expert System to compete with big names like IBM and provide the best technical solution to our customers. A deep knowledge of the sector and the reputation we have earned among our customers complete the cocktail.

LT-Innovate: What are the main new technology/linguistic challenges (recent and looking forward)?

Daniele Cordioli: In addition to the challenges mentioned earlier in the conversation (cross-modality, cross-channel, customer interaction and regulations), today’s global world will increasingly require cross-lingual analytics that support the customer regardless the language, enabling rich interactions and addressing complex situations. This will be key particularly for the European economic space, which is incredibly fragmented in terms of language if we compare it to other economies of similar scale.
Furthermore, language technologies have the challenge to merge recent progress in Artificial Intelligence, knowledge representation and reasoning to be increasingly deep, scalable and flexible. We have seen that structured symbolic approaches based on structured knowledge graphs with millions of explicitly represented concepts and relations can produce a deep understanding of the text, similar to the way humans read. However, they can sometimes be brittle in terms of the different ways in which knowledge can be expressed. On the other hand, Machine Learning has proven to be a cheap and fast way to address similar goals, based on learning mathematical functions that predict how data are correlated. However, though Machine Learning is good at identifying statistical correlation in the data it lacks the capability to produce logical deductions and explain the outcomes of such reasoning in ways understandable by humans. In the next years, we will see an increasing amount of work in academia and industry merging the symbolic and structured approaches for the greater good. We are already doing it at Expert System.

LT-Innovate: What is likely to be the most impactful change in this segment in the next 5 years?

Daniele Cordioli: In addition to the continued development of language technologies, we would highlight blockchain as a deeply transformative force in banking, which will require the key players in the sector to approach the business from a different angle. Similarly, the combination of robotic process automation (RPA) and cognitive technologies is due to revolutionize the insurance market.