Making credit “discoverable”: How to position your lending products for an agent-first economy

Interview with Yaacov Martin, CEO of Jifiti

With today’s higher rates and economic uncertainty, where are financial institutions most exposed when it comes to credit risk management and liquidity, and where do you see risk being misunderstood? 

Yaacov: The biggest exposure isn’t only credit losses, but misalignment. In uncertain periods, consumers and small businesses don’t stop needing credit, instead they become more sensitive to timing and relevance. When institutions respond with blanket tightening or generic offers, they push customers towards the most convenient options, which is often credit cards, whether that’s the best fit for them or not. That can increase risk for both the borrower and the bank.  

Liquidity strategy also gets misunderstood as ‘pull back and preserve.’ That’s one lever, but the more durable approach is precision: offering the right product for the right use case, at the right moment, with the right controls. In times of economic stress, the goal needs to be smarter credit that protects liquidity while keeping the institution present when customers need stability most. 

Small-dollar loans are often most in demand during periods of economic uncertainty, yet many community banks and credit unions still struggle to offer them. Why has this small-dollar lending gap persisted from a risk and cost perspective? 

Yaacov: Most small-dollar loan demand comes from everyday, banked customers dealing with everyday needs, like unexpected expenses or short-term cash gaps. Yet the issue stems from economics and process. 

Traditional lending infrastructure wasn’t designed for small lending at scale. When origination, verification, and compliance steps are manual and fragmented the cost to issue a $500 or $1,000 loan can cost just as much as a larger amount. That makes the math hard to justify, even when the borrower is low risk. 

When smaller institutions have access to modern digital lending infrastructure that’s affordable, the economics change. The cost to serve drops, small-dollar loans become viable, and underwriting standards don’t have to be compromised. For community banks and credit unions, that opens the door to meeting customer needs while keeping both risk and relationships in check.

What role does modern digital lending infrastructure play in reducing cost-to-serve and improving liquidity predictability, particularly for smaller banks and credit unions? 

For smaller institutions, lending can feel unpredictable due to manual reviews, disconnected compliance checks, and slow disbursement, lengthening timelines and making outcomes harder to anticipate. That lack of consistency makes funding and liquidity planning more difficult than it needs to be.

Modern digital lending infrastructure reduces that variability. Automated workflows and real-time verification bring consistency to decision cycles, making outcomes easier to forecast and costs easier to control. When banks know how quickly loans convert and when funds are drawn, they can plan liquidity more accurately, without pulling back from serving customer needs. 

There’s also a direct cost impact. Digital platforms can significantly reduce origination and servicing expenses, which changes the economics of products that were previously impractical, like small-dollar loans. Just as importantly, they keep customers inside the bank’s ecosystem, resulting in lending that’s efficient, and more stable from a liquidity and relationship standpoint. 

How does AI-driven decision intelligence help banks manage credit risk during periods of macroeconomic uncertainty? 

Yaacov: The most practical impact of decision intelligence shows up in how decisions are made, routed, and surfaced. That includes underwriting and compliance, and extends to discoverability, with how credit options are found and selected in the moment of need. 

As AI-driven tools and agents play a larger role in exploring financing, lending products need to be structured in ways that machines can understand and interpret. If an offer isn’t machine-readable or accessible through APIs, it may never surface at all, regardless of how competitive it is. That has serious implications for risk, relevance and market presence. 

When economic conditions change quickly, static rules and slow decision cycles struggle to keep up. This is when decision intelligence matters. It allows institutions to adapt in real time, matching the right product to the right use case, applying an appropriate level of verification, and delivering funds without unnecessary friction. In uncertain conditions, that ability to stay precise and responsive becomes a type of risk management itself. 

What common mistakes do you see banks make when they apply AI to credit risk management without proper governance and oversight? 

Yaacov: One common mistake is treating AI as an add-on while leaving fragmented processes underneath. Models may be sophisticated, but if decisions can’t be audited or explained, institutions create new forms of risk. 

Another problem is overlooking how AI reshapes access. As lending becomes more influenced by AI-driven discovery and automation, decisions are increasingly shaped by what systems can process and see. Without proper governance, this can introduce unintended bias. For example, products are favored because they’re easier for machines to interpret, not because they’re necessarily a better fit for a customer. 

Strong governance means designing workflows with transparency from the start: clear audit trails, defined escalation points, and human oversight where judgement matters. AI should make lending more accountable, not harder to understand. When governance is built into the decision framework, institutions gain speed while retaining control, especially when conditions change quickly.