Since its introduction in November 2022, OpenAI’s text-generating AI chatbot, ChatGPT, has been making headlines as a transformative technology that enhances efficiency and productivity.
The latest FIS Global Innovation Report reveals that generative AI holds the most significant growth potential in the coming years, with 92% of firms expecting adoption within the next 12 months. Furthermore, among the firms in Singapore currently using this technology, 67% plan to increase their investment in generative AI over the next year.
Singapore’s three largest banks — DBS, OCBC, and UOB — are also exploring and leveraging this technology to enhance employee productivity through the introduction of generative AI-powered chatbots and productivity tools.
How, then, can ChatGPT, and similar generative AI tools benefit the commercial lending sector?
The story so far
AI’s previous iterations have significantly contributed to the automation of commercial lending processes. Machine-learning AI, in particular, excels at extracting the right data from credit applications for digital processing or making relatively simple lending decisions.
As these tools continuously mine more data points and learn from human behaviour, they become more accurate and efficient.
In recent months, generative AI has advanced several steps further. Rather than merely automating manual tasks and decisions, tools like ChatGPT dig even deeper into available data to create their own textual or visual content.
Suddenly, it’s possible to automatically produce comprehensive, persuasively written dissertations or disturbingly realistic deepfake images, positioning generative AI as both astonishingly clever and a potentially dangerous tool for cheating the system.
Valid concerns
So, while generative AI promises to revolutionise financial services, including commercial lending, it understandably raises concerns within the industry.
Firstly, it presents more sophisticated opportunities for fraud. If generative AI can produce initially convincing yet ultimately phoney images, such as those of Donald Trump’s arrest, what ‘s to stop it from fabricating documents to show a lending applicant is more profitable and creditworthy than they are, or even to suggest the existence of entities that don’t exist?
Within financial institutions, the emergence of generative AI tools is igniting new concerns. Major Wall Street banks, including Citigroup, Goldman Sachs, and JPMorgan, have reportedly banned the internal use of ChatGPT as they evaluate risks related to data privacy, cybersecurity, and system access.
Moreover, there’s the ever-present concern that AI tools might displace an increasing number of human jobs. This anxiety is amplified by generative AI, which threatens not only to automate routine tasks but also to encroach upon roles requiring creativity and cognition.
Rewarding opportunities
But once the industry has addressed these concerns, I think lenders should not overlook the many potential benefits of generative AI for their processes and personnel.
Positive use cases for generative AI in commercial lending could include:
- Know your customer: Elevating problem identification to a new level.
- Credit assessment: Evaluating the creditworthiness of new businesses lacking a credit history.
- Fraud detection: Converting unstructured data into actionable insights, ensuring no warning sign is overlooked.
- Product generation: Analysing vast customer data to create highly tailored solutions and recommending them at optimal times in the business cycle.
- Credit application feedback: Providing customers with constructive insights on lending decisions.
- Financial analysis and forecasting: Anticipating future scenarios for customers or markets.
- Report generation: Crafting more user-friendly reports and dashboards, customised for the audience’s understanding.
- Model training and validation: Aiding in the creation of stress test scenarios.
- Sentiment analysis: Deciphering data from news feeds, social media, and other digital channels on businesses and sectors.
- Assisted credit memos: Supplying comprehensive background information for manual analysis.
What about human intelligence?
Like other forms of AI-driven automation, generative AI can remove friction from a wide range of commercial lending processes by processing vast quantities of data more swiftly, efficiently, and accurately than humans ever could. Crucially, it can also remove human bias and help lenders make impartial, thoroughly informed decisions.
However, AI tools lack empathy, a critical quality in traditional lending. Sometimes, when the financials are inconclusive, a strong gut feeling, or a profound understanding of the customer can kick in and lead to a successful deal.
Ultimately, then, while AI enhances human decision-making, it is not a substitute. Emotional intelligence and human insight remain crucial in credit assessment and loan management, working alongside AI tools.
But change is imminent and accelerating. For generative AI to fulfil its potential as a force for good, commercial lenders must quickly find ways for it to work in harmony with their human employees.
