Credit Underwriting in the Digital Era: Alternative Data and AI-Based Credit Scoring
Gaia Cioci - Head of Corporate Business Development at CRIF
Criff, an Italian company with global subsidiaries, leverages alternative and internal transaction data, augmented by artificial intelligence (AI), to revolutionize credit underwriting for banks, telcos, and utilities. By analyzing 50 million transactions and employing AI-driven categorization engines, they improved small business credit risk management by 14% and reduced cost of risk by 10%. AI also enhances client onboarding, document analysis, and personalized support, streamlining processes and standardizing evaluations. Generative AI promises future efficiency gains, contingent on organizational adaptation and open banking data integration.
"To experience credit underwriting in the digital area, we must focus on data—especially alternative data—and artificial intelligence, which allows for more accurate analysis, speeds up processes, and transforms traditional methods into efficient solutions."
Summary
- Credit underwriting growth expected in mortgages, consumer, and small business lending in 2025 and beyond. - Key gaps include manual data handling during onboarding, lifecycle, and credit evaluation processes. - AI and transaction data improve credit risk management and reduce cost of risk significantly. - AI-driven platforms enable better cross-selling and loan proposals, especially for small businesses. - Combining document analysis, data extraction, and content creation boosts precision, standardization, and analyst efficiency.
Article
AI and alternative data transforming credit decisions, Nordic Fintech Summit told
Italian credit specialist reveals how transaction data and artificial intelligence are revolutionising lending to small businesses
The future of credit assessment will be driven by alternative data sources and artificial intelligence, with significant benefits for both lenders and small businesses, according to a leading industry expert speaking at the Nordic Fintech Summit in May.
Gaia Cioci, head of corporate business development at Italian credit information provider CRIF, told attendees that traditional credit underwriting processes continue to suffer from significant gaps that new technologies can address, particularly for small businesses where conventional financial data is often limited.
"Artificial intelligence gives us the possibility to analyse information more accurately and speed up the credit evaluation process," Cioci said during her presentation on the opening day of the two-day event.
Leveraging internal data treasures
While many financial institutions look externally for alternative data sources, Cioci emphasised that valuable insights often already exist within their systems. Transaction data – the record of customers' financial activities – remains an untapped resource in many organisations.
"Transaction data is internal and often unexplored data that can be categorised and analysed by AI to provide precise creditworthiness assessments," she explained.
CRIF's approach has involved developing AI-powered categorisation engines that analyse transaction patterns and generate key performance indicators specifically designed for credit assessment. In one case study presented at the summit, the company processed 50 million transactions and created 100 distinct categories to enhance lending decisions.
"By analysing fifty million transactions with AI, we increased preventive credit risk management by fourteen percent and reduced cost of risk by ten percent," Cioci told attendees.
Beyond risk: enhancing customer relationships
The benefits of AI-driven credit assessment extend beyond risk management, according to Cioci. The same technological approach enables financial institutions to better understand customer needs and provide more timely and relevant services.
"Using transaction data and AI to offer the right loan and service at the right moment led to a twenty-four percent increase in cross-sell for small business clients," she said, highlighting how enhanced understanding of customer behaviour creates commercial opportunities.
This customer-centric approach appears to be welcomed by clients themselves. Cioci noted that almost all small businesses in their case studies willingly provided consent for their transactional data to be accessed via open banking connections, suggesting significant customer acceptance of data-sharing when it delivers tangible benefits.
The rise of generative AI
The presentation also addressed the emerging role of generative AI in credit processes. CRIF has developed applications that combine document analysis, data extraction and content creation to produce standardised credit assessment reports.
Rather than replacing human analysts, these tools provide an enhanced starting point that ensures complete documentation review while saving time and increasing precision. This ensures consistent evaluation standards across an organisation, Cioci explained.
However, she cautioned against expecting immediate transformation from newer AI technologies: "Generative AI won't instantly boost efficiency; it requires one to two years of change management to achieve its full potential in processes."
Future outlook
Looking ahead, Cioci predicted continued growth in mortgage, consumer and small business lending throughout 2025 and beyond, with AI and alternative data playing increasingly important roles in credit assessment.
The summit, which continues through 15 May, brings together fintech leaders from across the Nordic region to explore technological innovation in the financial sector.
For financial institutions still relying on manual data handling and paper-based processes, Cioci's message was clear: the future of credit underwriting lies in digital transformation, with those who embrace AI-driven assessment gaining significant advantages in risk management, operational efficiency and customer satisfaction.
Part of Nordic Fintech Summit