AI, Open banking, and the future of SME lending: Key takeaways from FTT Lending

Lending is evolving, but perhaps not as quickly as you might expect. At last month's FTT Lending event, industry leaders came together to discuss the ongoing transformation of business (SME) and consumer credit markets. From Open Banking to AI-powered decisioning, these discussions offered valuable insights into how fintechs and lenders are adapting to change. This blog explores the key takeaways from those conversations.

SME lending remains a massive market, valued at around $250 billion according to Allica Bank. However, we were told that while $5 billion of this is unsecured and ripe for digitalisation, the majority is secured against tangible assets, making it harder to modernize. Today, many processes remain frustratingly offline, with wet signatures, site visits, and manual checks still the norm.

Despite the rise of digital lenders, Alliance Bank also stated that traditional broker-originated loans still make up a significant share, accounting for $38 billion of SME lending annually. There are around 60 active asset finance companies, and while fintechs are driving change, the industry has a long way to go before true digital transformation is achieved.

With modern lending, Open Banking can be a critical tool and may be helpful to better serve SMEs and underserved consumer markets, particularly for SMEs and underserved consumer markets. Many traditional credit scoring models rely on data from the three main bureaus, but this information is often 4-8 weeks out of date. Open Banking can offer real-time access to financial data, making lending decisions more accurate, faster, and cost-effective.

At the event, firms like Jaja Finance and Allica Bank highlighted that by incorporating Open Banking they may be able to improve their credit models and offer better lending terms. Panel experts explained why they see Open Banking not only speeding up account access but also lowering costs for lenders by enabling them to provide instant, permissioned access to financial histories. However, while fintechs like Funding Circle, Iwoca, and Capital on Tap have embraced Open Banking, many SMEs, perhaps surprisingly,  still rely on personal credit cards to fund their businesses.

We were told that for some lenders, Open Banking adoption has been slower due to alternative access to current account data. Allica Bank told us that while this provides a snapshot of an SME’s recent financial health, it lacks the granularity of transaction-level data, limiting its predictive power, but serves lenders pretty well nonetheless. Open Banking is proving invaluable for thin-file borrowers, self-employed individuals, and those new to a country, helping them secure financing where traditional credit scoring might result in rejection.

AI was a dominant theme throughout the event, but the consensus was clear: AI should enhance human decision-making, not replace it.

Lenders like Jaja Finance are leveraging AI and machine learning to refine credit models, enabling faster, more personalized lending decisions. Instead of relying solely on static credit scores, AI-powered models update every six months, continuously improving based on new data. This results in quicker approvals, more tailored loan offers, and better risk assessment.

In customer service, AI-driven chatbots are reducing response times dramatically. For example, Jaja's large language model (LLM)-based chatbot handles 50% of customer inquiries, responding in just 9 seconds compared to the 3-minute average for a human agent. This is a game-changer for scaling customer support while maintaining service quality.

AI is also proving valuable in fraud prevention. Persistent.ai’s fraud detection tool screens documents for manipulation, identifying around £1 million in fraudulent applications every week for one SME lender in attendance. Similarly, AI is being used to scan contracts, extracting key terms and flagging high-risk clauses - an essential tool for risk management in lending.

While AI is playing an increasing role in lending, the industry remains cautious about full automation, particularly in complex decision-making. The consensus at FTT Lending was that AI should assist human underwriters by providing better data and insights, rather than making final lending decisions.

For straightforward applications, AI-driven decisioning is already a reality. But when it comes to complex cases, such as high-value SME loans or unique borrower circumstances, humans still need to be in the loop. AI can help lenders say "yes" more often by streamlining document analysis and risk assessment, but final approval remains a human task. Interestingly, some brokers actually prefer a bit of friction in the process, as it provides a level of reassurance that due diligence is being conducted properly.

Looking ahead, AI is expected to play an even greater role in shaping lending strategies. Rather than relying on rigid, pre-set rules, the future of AI in lending lies in natural language rule creation, according to GDS link. In this model, lenders could ask AI to analyze the impact of adjusting risk parameters, such as increasing loan limits or tweaking affordability criteria, and receive real-time predictions on how this would affect lending volume and risk exposure.

Panel host and MD at BPG Strategy, Brendan Gilmore, made an insightful comment: "Banks want to be first to be third." Many large institutions are hesitant to be the first movers in AI and Open Banking but are quick to follow once competitors prove the model works. As AI-driven lending matures, expect more incumbents to jump on board.

The lending industry is at an inflection point. AI, Open Banking, and digitalization are reshaping how credit is assessed, approved, and managed. However, the journey is far from over. SME lending still faces significant barriers to full digitalization, and while fintechs are pushing the boundaries, traditional processes remain entrenched.

The key takeaway? AI and Open Banking are enablers, not disruptors. Lenders that embrace these technologies will unlock faster, more efficient lending while maintaining the human touch that borrowers value. Whether it’s real-time credit modeling, AI-powered fraud prevention, or automated customer service, the future of lending is about augmentation and not replacement.

For fintechs, lenders, and banks alike, the message is clear: evolve, innovate, and adapt. Or risk being left behind.

Marqeta powers some of the most innovative companies, who leverage our open APIs, real-time transactional data, and web push provisioning capability to help create faster, smarter, personalized lending experiences.