The financial services industry has begun to undergo a significant transformation following the introduction and application of artificial intelligence (AI) over the last several years, says George Roth, CEO, Recognos
While the industry remains heavily data-dependent, about 80% of the underlying data being processed remains either semi-structured or not structured, and is being processed using costly, inefficient and risk intensive manual processes. Artificial intelligence applications are changing these practices, making it simpler and more profitable for firms to analyze and contextualize their data. The most common AI categories being applied in financial services include natural language processing (NLP), data mining and text analytics, semantic technologies (ontologies, linked data, sentiment analysis) and machine learning.
In order to be able to access enterprise data as a ‘whole’, natural language-processing techniques are used to prepare the data for downstream AI-related analysis processes. It will become a mandatory ‘infrastructure component’, which when combined with semantic technology will make the difference between ‘structured’ and ‘unstructured’ disappear in data processing. In future, the following are the areas where we see financial institutions applying AI related technologies:
Generalization of the semantic search: Search will be transformed into a performant tool to access information in a holistic way. Search will evolve from the ‘keyword search’ approach to a natural language-based search, where the user can address a question using the natural language. The trend can already be seen in the evolution of Google and the migration towards a question / answer solution. Answers to search questions will be provided from completely integrated databases that contain structured and semantically tagged data. In order to achieve these, unstructured documents will be semantically tagged using NLP, ontologies will be built to support the search and the linked data will be used for enhancing the search results with information available inside and outside the enterprise.
Lead generation and lead management: Using AI technologies, financial services companies will be able to find leads from different public sources, qualify leads and optimize the sales process by targeting prospects with very efficient and less intrusive campaigns. The key for this is gathering information from multiple sources, integrating them and enhancing the information by using linked data. For example, an investment company specializing in high-net-worth individuals will be able to detect large purchases in certain areas of interest and determine the approximate net worth of the buyers. They can then cross-check to determine if these individuals are in their lead or customer databases (this is not a trivial problem, and also can be simplified by leveraging semantic de- duplication techniques and generate a new lead. In this way, internal ‘lead factories’ can be generated.
Compliance, fraud detection, security: By integrating structured and unstructured data, AI techniques will help financial services companies increase the efficiency of enforcing compliance rules (e.g. in customer communication), prevent fraud (insider trading events), detect money laundering schemes and detect and prevent intruders from access into secure networks. A fundamental property of these solutions is that all these systems will auto-develop in time based on the detected incidents.
Currently, business reporting is based on analyzing current data that resides in data warehouses and traditional relational databases. These reports can only solve problems that ‘we know that we don’t know’ – for example finding in time how many clients had a set of pre-defined different properties. The big leap using AI in this area is to be able to find these things, i.e. discovering new facts and building predictive models using the technology bundle known under the name of “data analytics”. This new reporting paradigm will allow financial institutions to create reports on the fly and to integrate data from multiple sources, creating a more clear and accurate picture to improve business performance.
Customer relationship optimization: By understanding enterprise data as a whole and merging information contained in structured and unstructured data silos, customers will become better known. This will allow companies to implement predictive customer relationship actions, minimize customer complaints or to quickly fix issues that can cause a high percentage of churn in the enterprise.
There are many other areas in financial services enterprises that will benefit from the AI technologies and ultimately will allow for continual business improvements, higher qualities of service and increased customer satisfaction. These AI technologies, which were often thought of as “rocket science” in the past, will become a must for both buy and sell side financial firms in order to create and maintain a competitive advantage in the market.