How we built a credit model from scratch without credit scores

AI-driven underwriting is rewriting the rules for lenders—are you ready for what’s coming in 2026? Every week, I hear from fintech and lending leaders anxious about staying ahead of seismic shifts in risk assessment and compliance. The good news? The future is bright for those who adapt early. Here are 3 key AI underwriting trends every lender should be watching for 2026: 1️⃣ Contextual Decisioning: Traditional scorecards are out. Modern AI underwriters will weigh nuanced borrower behaviors, macroeconomic signals, and real-time data to power safer, faster approvals. 2️⃣ Explainable AI as Standard: Regulators and boards demand transparency. Black-box models will fade—tomorrow’s solutions must offer clear, audit-ready decision paths. 3️⃣ Proactive Fraud Intelligence: AI will spot synthetic fraud and risk patterns upfront, not just after the fact. Real-time portfolio monitoring could become a competitive edge. I’m fascinated by how these trends unlock smarter scaling, compliance peace of mind, and meaningful borrower relationships. The question: how will your lending platform keep pace? What trend are you most excited (or challenged) by? Drop your thoughts below—or DM me if you’re exploring ways to upgrade your underwriting for the next wave of innovation. Let’s future-proof lending, together. #Fintech #Underwriting #AIinFinance #LendingInnovation #FinancialServices

What’s the real cost of slowing down your lending process—and who feels it most? For growing banks, every delay means lost trust, missed opportunities, and frustrated teams stuck doing busywork. That’s why Vantage Bank Texas, Cavallo Technologies Inc., and Databricks joined forces to build something different: an AI lending assistant that doesn't just streamline workflows—it transforms them. Imagine 94% accuracy on complex lending questions. Approvals occur faster, errors are eliminated, and lenders finally have time to focus on what matters: building relationships. Check out the full story and join the conversation: https://lnkd.in/gwrnMGY4 Shawn Main Guillaume Colley Rob Cavallo Antoine Amend Kimberly Hatton Junta N. Jennifer Miller Rajaram Suresh Jamie Cosgrove Erin Butler Marcela Granados Lavoie #AI #FinTech #Innovation #LoanProcessing #Databricks #CavalloTechnologies #VantageBankTexas

𝗔𝗜 𝗶𝗻 𝗧𝗿𝗮𝗱𝗙𝗶: 𝗠𝗮𝗸𝗶𝗻𝗴 𝗖𝗿𝗲𝗱𝗶𝘁 𝗦𝗰𝗼𝗿𝗶𝗻𝗴 𝗠𝗼𝗿𝗲 𝗜𝗻𝗰𝗹𝘂𝘀𝗶𝘃𝗲 AI is revolutionizing credit scoring, moving us far beyond old-school, rule-based systems (RBC) that relied only on things like BVN, income, and collateral. It worked, but it was slow and left out a lot of people who didn’t have a strong banking history. Now, with AI-driven credit scoring, things are moving faster. AI can look at how people pay their bills, buy airtime, or shop online. These small daily actions can now build a kind of “micro” credit history. This is making lending faster and more accessible for many. But here’s the challenge: what happens to people who don’t have any digital record at all? Many still live mostly in cash economies, without smartphones or online transactions. To AI, these people are invisible and that means they risk being excluded from access to credit. So how do we include them? One answer lies in agency banking and community data The Path to True Inclusion: Community-Based AI Agents who process cash-in/cash-out, deposits, and airtime top-ups hold valuable, real-world data. By applying AI to these agent networks, we can develop community-level risk models. This allows AI to learn from the systems and communities that serve people, assessing risk collectively when individual digital trails are nonexistent. Using AI for community-based credit scoring can sound complex, but it doesn’t have to be out of reach. The cost and resource needs really depend on how it’s implemented. By starting small, using existing agent network data, and leveraging open-source or cloud-based AI tools, financial institutions can build inclusive, data-driven credit systems without massive budgets. This approach transforms the AI's learning from individual data points to systemic inclusion. The Ask: How Else Can We Close This Gap? Beyond leveraging agency banking data and community risk modeling, what other non-digital, ethical data sources can AI leverage to fairly assess those without a formal digital footprint? #FinTech #AI #FinancialInclusion #TradFi #CreditScoring #AgencyBanking #CommunityData #Innovation

Can AI Help You Improve Your Credit Score? Apps like CRED already know our spending and repayment patterns. The next evolution is using that data intelligently to help users build healthier credit profiles. Imagine AI that: 🔹 Warns when you’re nearing 80% of your credit limit 🔹 Suggests switching cards to balance utilization 🔹 Recommends optimal repayment timing to boost your score 🔹 Flags risky spending behavior before it hurts your profile Fintechs have enough data to not just simplify payments — but to coach users toward better credit health. The future of personal finance isn’t just AI-driven automation. It’s AI-driven financial discipline. #AI #FinTech #MarTech

It’s been almost a month since Finovate and I’m still thinking about some of the demos I saw there. Not because they were flashy, but because they showed real solutions to real problems in financial services. Things that can change how institutions run, how compliance is handled, and how people experience money. Here are 7 that stood out and stuck with me: Warrant: AI agents for marketing compliance. Reviewing, tracking, and record-keeping, all automated. Compliance teams get speed and trust. (Great session by Austin Carroll) Appli, Inc: making banking sites smarter. AI-powered recommendations that transform calculators into real competitive advantages. (Tim P. App showed this with clarity) Castellum.AI: end-to-end compliance with AI agents + global risk data. False positives drop, compliance confidence goes up. (Peter Piatetsky led a strong demo) Vertice AI: growth for community institutions. Turning data into action — personalized recommendations without a data science team. (Mitch Rutledge nailed the value) Gentreo: beyond documents. Helping organizations support employees and customers through life’s biggest moments, with legal and digital vault solutions. (Julie Fry & Renee Fry with an authentic presentation) Krida: commercial lending without the drag. Automating pre-underwriting, doc collection, and tracking so banks close faster and borrowers stay engaged. (Sophie Jewsbury with a sharp walkthrough) WNSTN AI: next-gen investment management. Personalized portfolios, real-time insights, directly in broker platforms. (Roy M. & Jamie Rakover showing the future) A month later, these are still on my mind, proof that when fintech demos solve real pain points, they stick. #Finovate #Fintech #AI #FinancialServices #Innovation

Most credit unions are leaving growth on the table. They're using AI for underwriting—but stopping there. Meanwhile, opportunities to deepen member relationships, predict needs, and prevent churn are slipping through the cracks. The difference? Thinking beyond approval. AI's real power isn't just saying "yes" or "no" faster. It's transforming the entire member journey—from the first prescreen to cross-sell opportunities to collections. It's turning data into action and fragmented touchpoints into seamless experiences. Join Jeff DeShane (VP of Consumer Lending, People Driven Credit Union), Tracy Kaufman (Chief Lending Officer, Wildfire Credit Union) and Eric Steinhoff (EVP Client Impact, Scienaptic AI) in an on-demand webinar in partnership with MeridianLink, as they break down how to: ✓ Grow approvals without inflating risk ✓ Personalize experiences at every stage ✓ Build the data infrastructure that makes smarter decisions possible This isn't theory. It's how leading credit unions are winning. Webinar Link - https://lnkd.in/eSfRy8wV #CreditUnions #AI #Lending #MemberExperience #Credit #FinancialInclusion

The #1 Mistake Banks / Credit Unions Make while Launching AI Starting with the 'How' instead of the 'Why.' It's common for lenders to say, "We want to use more AI – where can we deploy it?" This 'solution-first' approach often leads to fragmented efforts and unclear ROI. The more effective strategy? Work backward from the business problem. Instead of pushing AI everywhere, focus on core goals like:  "How can we approve more loans in a way that creates value and brings down our default rates?" Once the 'WHY' is crystal clear, AI becomes a powerful tool to achieve those specific outcomes – not a technology looking for a home. This problem-centric view makes it much easier to organize teams, align stakeholders, and ensure your AI initiatives deliver real, measurable impact. Eager to hear more WHYs at #CUReach2025 event tomorrow with California's Credit Unions and Nevada's Credit Unions #manispeaksmoney #CreditDecisioning #BankingInnovation #Banks #CreditUnions #FinTech #DigitalLending

This was a fun conversation with Jeannette (Robinson) Friedrich on how we're integrating AI Agents into our lending business, Zendra Labs. Specifically, we built a voice & chat agent that replaces intake forms with a conversational AI. Everybody knows the tech is getting better quickly, but you may be surprised (or not) to hear that people still want to talk to people. There's a lot of hesitancy from borrowers in working with AI voice and chat agents; in most cases they'd rather work directly with a human. In fact, all $8M in loan originations we've done to date have been through warm relationships & manual intake on details. Even so, AI in lending is like a pizza shop with a website 2005. We're early adopters, but this is going to be table stakes in the future. Listen to the full episode in the comments. Blue Lake Capital Burns Capital Partners

🚦 Explainable Velocity: The Real Test of 60-Second Loans We all love the headline: “Loan Approved in 60 Seconds.” But here’s the uncomfortable truth — speed without explainability is a liability. In today’s AI-first lending world, speed alone doesn’t impress regulators or customers. They want “explainable velocity” - every decision traceable to its logic, data, and reasoning. Let’s decode it 👇 ✅ Real-time data retrieval - every click triggers an event stream ✅ Automated document intelligence - no manual file checks ✅ AI credit scoring - fast, but bounded by rule-based guardrails ✅ Full audit trails - every model decision logged and explainable Banks are realizing: it’s not about approving faster; it’s about justifying faster. “Speed that can’t explain itself won’t survive scrutiny.” That’s the real evolution of lending: from faster systems to trustworthy systems. #AIinLending #ExplainableAI #FintechInnovation #AgenticAI

In a new FintecBuzz Q&A, our CEO Philip Paul shares how AI Growth Agents help community banks and credit unions: 🔶 Boost digital growth and personalization in real-time 🔶 Convert single-product members into multi-product loyalists 🔶 Automate decisions—like account openings and loan approvals—in under 2 minutes 🔶 Leverage AI without replacing legacy systems Philip also dives into how institutions can compete with megabanks and fintechs by pairing automation with human creativity, turning onboarding into a growth engine. 🎧 Read the full Q&A here: https://lnkd.in/dkVnnzbg #CommunityBanking #Fintech #DigitalTransformation #AI #CustomerExperience #Growth #Cotribute