May 18, 2017

Changing customer behavior, rising competition, and eroding loyalty make it imperative for financial institutions to deliver a differentiated experience, contextualized to customers. Enter cognitive computing; the journey toward ensuring customer centricity hinges on a banks ability to build a cognitive system using customer information. Such systems have the ability to listen, understand, infer, adapt, correlate, and make decisions.

Cognitive computing can further help banks initiate proactive engagement with customers.

Cognitive Computing

Cognitive Science is an interdisciplinary study about the workings of the nervous system that represent, process and transform information. Applying cognitive science in the computing field has led to the development of cognitive computing technology. Broadly, while AI focuses on all emerging technologies that gather intelligence, cognitive computing is a sub field under AI, where the focus of intelligence gathering is based on learning from processes of the human brain.

Banks can incorporate natural language processing (NLP) and human-computer interaction (HCI), along with machine learning techniques, in self-service and assisted channels such as chatbots.

Real time Contextualization

Banks are leveraging NLP techniques to digitize communication across the spectrum from customer interactions, internal collaboration, to communication with potential hires. NLP techniques use retrieval-based mechanisms that provide a structured way of extracting intent from such communications. Some players such as TensorFlow have brought in word embedding techniques to effectively model the language and learn the features essential for improving the accuracy of language understanding.

Even as the banking and financial services industry focuses on these areas, it is important to provide effective responses to contextual customer queries in real time. Much like the hippocampus the part in the human brain that is associated with memory and spatial navigation — the cognitive system will help banks correlate contextual knowledge to real-time situations, which essentially means recalling and aligning information relevant to an event. This will help banks capitalize on cross-sell and upsell opportunities, as well as identify and address issues proactively. For real-time contextualization to work, it requires effective knowledge representation, a common meaning and linkage across inter-connected entities (or systems) in real-time, as well as efficient computational procedures.

For instance, people tend to overspend while on vacation. By combining location data, account data, and historical information about spend patterns and customer behavior, banks can proactively warn customers about the likelihood of their accounts getting overdrawn. And at that crucial moment, banks can even offer to discount or completely waive overdraft charges. Getting such an offer at the right moment could well build lifelong loyalty. Similarly, customers navigating the banks website can be informed of the right product for their needs based on current context and browsing patterns combined with historical knowledge. In all such cases, the hippocampus built by the bank will track customer behavior in real-time with respect to the location, mobile, online activity, and so on. This information is then connected with relevant knowledge such as life stage, brand affinity, location preferences, and financial scenarios to identify the right context, and come up with the right offer.

A cognitive brain is the pivot of customer-centric banking enterprises. Building a cognitive computing platform at the back-end, which has the capability to recognize hidden patterns in data flowing in from disparate sources in real-time, using effective computational algorithms and knowledge representation techniques, is important. This will help deliver differential customer experience. With fast data streaming, NLP, machine learning, and knowledge representation techniques attaining a certain degree of maturity, the possibility of creating a foundation around which banks can build a cognitive brain is fast becoming a reality.


Navin MK is a senior consultant and banking and financial services solution architect at Tata Consultancy Services, India. He co-owns the offering on the Customer Knowledge Graph and the Real-Time Contextualization space. His experience covers retail and mortgage banking, commercial lending, regulatory risk, compliance and investment banking. His research interests are in the areas of fast data streaming, cognitive intelligence and semantic knowledge representations.