GenAI, AI & Data: towards a customer-centric finance

Opportunities and challenges in the evolution of the financial industry

Artificial intelligence, GenAI and the power of data are revolutionizing the financial industry.

GenAI, in particular, is the latest evolutionary step in a technology that has been radically helping to automate operational processes and analytical tasks for more than 20 years.

If machine learning, the first step of AI, applies models and algorithms to available and, usually, structured, proprietary and specific databases to identify commonalities and make predictions, GenAI goes a step further: it typically relies on foundation models trained on huge amounts of data, such as language, images, sounds, music. These models are able to learn the distribution of data and have developed a high capacity to read, write text, generate new content and chat with humans.

Great Expectations

The expected impact is enormous, ranging from the reduction of costs for human-intensive activities (from coding to writing any kind of document or basic analytical tasks) to the possibility of redefining everything that has to do with supporting or relating to people, ‘translating’ data and information into structured and complex dialogues.

This world is still in its early stages, even in the financial sector, and still suffers from a strong hype. However, this will soon make way for more tangible initiatives, such as those that enhance the relationship between customers and companies, making the management of their needs more fluid and customized, thus revolutionizing the industry.

Banks and insurances are working on various initiatives of testing, experimentation, and also investment in the internal teams: some are focusing on the AI code generation, others are testing new chatbot models to reengineer call centers, and others are working on supporting their consultants by converting reports or manuals into more user-friendly and intuitive systems.

GenAI has the potential not only to improve the customer experience by making interaction simpler and more intuitive, but also to reduce waiting times and lighten the workload on human support teams by optimizing resource management.

To give just one example in the banking sector, all the activities related to resolution processes, from ex-ante analyses to ex-post evaluations, can now be re-engineered, integrating within them language models capable of preparing drafts of the main documents and re-reading the notes produced by operators to verify that all company policies have been respected.

Early outcomes

In this context, during Tomorrow Speaks 2024 - a CRIF event dedicated to the transformative potential of AI and GenAI and data for the financial industry - we presented one of our real use cases: an application we developed with Generative AI for the appraisal and origination phase of corporate financing. This tool automates the analysis of business documents, such as balance sheets and property appraisals, turning manual KPI extraction into a precise and efficient process.

Another case in point is our innovative tool for second-level controls in banks, which automates the analysis of loan files, streamlining the process through automatic pre-evaluation that allows operators to process 100% of files.

From 2025 onwards, the challenge will be to move away from the experimental approach and towards an industrial use of generative, harnessing it to answer real and circumstantial questions.

In the short term, the expectation is that GenAI will have a major impact on operational costs (whether IT or back office and operations), as the issue of efficiency and productivity is an integral part of the business strategies of all companies with long-term perspectives.

Take care of “good old” AI

While exploring the potential of GenAI to build tools increasingly capable of speeding up processes and making the user and customer experience better, the finance industry can still expect a lot from ‘traditional’ AI.

Machine learning and predictive algorithms have long been an ally in assessing and monitoring the creditworthiness of customers, retail and corporate, and enhancing Environmental, Social and Governance (ESG) indicators.

At CRIF with machine learning techniques we have already developed advanced models to assess the creditworthiness of companies, the ability of individuals to sustain financing, predict the risk of customer churn, analyze the probability of insurance incidents and monitor the risk of fraud in specific transactions.

For example, we developed an automatic score of sustainability and ESG factors for each of the more than 5 million Italian companies, using more than 160 type of data from various sources, to generate the three main indicators (environmental, social and governance) and a final synthetic score.

Cases like this reveal the fundamental importance and role of data, unique and of quality, without which artificial intelligence could not exist.

Quality data matter

Progressive digitization, from the arrival of the web until today, has created a wealth of data that allows AI to generate unprecedented results. But without the right technologies, predictive models, algorithms and human skills, navigating this mare magnum of data would be impossible.

The key to winning the challenge, especially in the medium to long term, is therefore linked to the opportunities offered by the combination of Data and AI, a strategic lever not only to optimize risk management but also to anticipate the needs of households and businesses, creating value through increasingly customer-centric experiences and bringing concrete benefits in terms of profitability and the consolidation of customer trust.

CRIF, as a Data Company, was born on the data of its Centrale Rischi and works constantly to enrich its wealth of data and information. We have grown by investing in analytical capabilities and technology, developing our concept of CRIF Metadata - unique, quality data - which fits naturally with everything AI, traditional or generative.

It has been a progressive evolution. The abundance of quality data and information we have today, combined with 15 years of daily experience in artificial intelligence and the expertise of a team of more than 300 AI experts and more than 1,500 ICT professionals, allows us to position ourselves as pioneers also in GenAI.