How to Improve Underwriting Processes

Same issues we had in 2021. Gift letter unsigned. CD sent late. Case number pulled wrong. Everyone accepts this as "just mortgage." We shouldn’t. We’ve convinced ourselves that these problems are inevitable. That dealing with fires is just part of the job. But here's what we know: The goal isn't to manage problems better. The goal is to make each problem impossible to happen again. It’s easy to approach these problems through: • More training • Better communication • Stricter oversight All of these are band-aids. You train someone not to forget the gift letter signature. They forget anyway because humans forget when managing 40 loans with 800 moving pieces each. You communicate better about CD timing. Someone still misses it because they're drowning in urgent requests. This is what we tackled. Fix it systematically. Not through training. The goal is to build systems that make the mistake impossible: • Gift letters can't be uploaded without a signature verification step • CDs automatically trigger based on loan timeline, not human memory • Case numbers get validated at the point of entry, not weeks later in underwriting Fix it once. Never see it again. This isn't about being harder on people. This is about being smarter about systems. The truth is we’ve accepted dysfunction as normal for decades. We don’t have to anymore. Processors need to stop having to remember 800 tasks across 40 files. Loan officers need to stop chasing updates on problems that could have been caught weeks earlier. Underwriters need to stop cleaning up messes that were knowable and preventable. Right now, everyone is constantly fighting fires, which means no one has time to build the systems that would prevent the fires in the first place. That's what we're changing. We're not going to accept "that's just how mortgage works." We're going to ask: "How do we make this impossible to happen again?" The answer isn't more training. It's what we're committing to build together: Systems that catch problems before they become emergencies. Workflows that surface issues when they're calm and manageable, not when they're weekend disasters. Processes that work even when humans are tired, distracted, or overwhelmed. Here's what we know: Twenty years from now, if we don't change how we approach this, teams will still be dealing with unsigned gift letters, late CDs, and wrong case numbers. The same problems. Forever. We're going to fix them systematically. Once and for all.

I’ve worked with a lot of auto lenders trying to speed up time-to-fund. Almost every conversation starts the same way: “We’re exploring AI for underwriting.” Or… “We just need more underwriters.” Totally fair instinct. Volume is up. Margins are thin. Speed wins deals. But here’s what I’ve learned: underwriting isn’t slow because of underwriters. It’s slow because the preconditions for a decision are a mess. This is what I keep seeing behind the scenes: ✅ Docs scattered across different sources ✅ Underwriters checking KBB on one tab, DTI in another ✅ Review rules living outside that platform in notepads, PDFs, or memory ✅ 40%+ of apps requiring manual review ✅ Dealer-submitted data needing re-entry, rechecks ✅ Underwriters having to do manual math and manul flags ✅ No structured way to push back on missing or invalid uploads ✅ Apps stalling, no one owning next steps, and no system to track why And then the question: “Can AI fix this?” Not if your system still relies on underwriters doing data janitor work. The fastest platforms don’t just have better models. Because underwriters aren’t the problem. The system is. The workflow is. The best platforms remove the noise before underwriting ever happens. The biggest lift for time-to-fund comes from these places: 📌 Upfront Borrower Data Intake - Dealer mismatches auto-flagged on submission (income, employment, addresses, etc.) - Borrowers upload income docs and connect bank accounts via link, not over email - System pre-checks for missing fields and docs, so your team doesn’t have to 📌 Data Clean Up - Auto-calculated DTI, PTI, LTV, and average monthly income across varied paystubs (no more Excel) - One-screen document management (rename, reorg, rotate, reject) with structured notes and flagging so underwriters can push docs back with one click - Underwriting rules visible in-line, so no more notepads and sticky notes - Auto-validation rules for income logic to auto-structure details, avoiding manual reviews. 📌 Manual Review Flow - Flagging edge cases early so clean apps fly through. - Edge cases come with context already attached - Review queues grouped by issue type (e.g. doc mismatch, income variance) not left to underwriter triage I’ve come to believe this: Underwriting isn’t just a decision engine. It’s a real-time coordination layer between dealers, borrowers, docs, risk rules, and processors. If that coordination layer is broken, AI won’t save you, and your time to fund won’t decrease. Thoughts? Anyone else rethinking where underwriting speed really comes from? Would love to hear what’s working (or not) for you.👇 If you found this relevant, little plug for Praxent, if you want to solve these problems and upgrade your auto-finance systems, check out our auto-finance page  👉 https://lnkd.in/gA4xjM8X #AutoFinance #AutoFinanceNews #AI #AIUnderwriting #DealerFinancing

Reduced loan underwriting time by 50%   (Check out the 4 AI Agents that made it happen) Loan approvals happening in half the time! With improved accuracy and a better customer experience. Not a distant future – it's happening right now. We've just helped a European bank customer achieve these results using a team of specialized AI agents. The impact? Dramatic improvements in efficiency, compliance, and decision-making. Here's how we did it: 1: Loan Origination Agent -- Powered by Azure AI services using GPT-4 and GPT OMNI -- Streamlined application process with instant, customized checklists for complex loans -- Result: Reduced errors and sped up approvals for hundreds of daily applicants 2: Loan Underwriting Agent -- Utilizes RAG Azure AI Services and OpenAI ADA embeddings -- Retrieves key data from lending history, real-time market data, and regulatory guidelines -- Helps loan officers deliver accurate risk assessments on high-value loans Ensures compliance and improves decision-making 3: Loan Audit and Compliance Agent -- Fine-tuned with T5 and LoRa -- Continuously reviews past decisions and flags anomalies -- Keeps the bank compliant with evolving regulations -- Minimizes computational costs 4: Loan Self-Reflection and Optimizing Agent (my fav) -- Leverages Codex and Autogen -- Learns from past underwriting decisions -- Makes the entire process smarter and more efficient over time .................................................................................. The bank is already seeing tangible improvements in processing times and accuracy while maintaining robust compliance. Will share more detailed results after the quarter closes. Which of these AI Agents could have the biggest impact on your underwriting process? #AIinBanking #FinTech #AIAgents

AI-driven underwriting is reshaping lending economics, and surprisingly few have caught on yet. I've been reflecting on why credit decisioning, especially to SMBs, remains so manually intensive. At first glance, you'd think regulation or tech limitations hold things back, but the core bottleneck is actually human labor. Banks still rely heavily on manual processing—reviewing outdated financial statements and Dun & Bradstreet reports, and depending heavily on human judgment to catch subtle risk signals. This problem feels familiar to one I worked on at Nauto (AI software for driver safety). Our models had to detect every crash perfectly (zero misses). But if we optimized strictly for perfect recall, precision plummeted. We flagged too many false positives, slowing down our human reviewers. So we built a human-in-the-loop system where AI pre-highlighted events, shrinking human review time down to just five minutes. The hybrid AI-with-human-oversight solution was key to managing scale and efficiency without sacrificing accuracy. Banks face the same recall-precision dilemma with underwriting. Traditional financial metrics, which are manually prepared, months old, and often incomplete, mean underwriters either miss important signals or drown in excessive manual reviews. At Slope, our hunch was that raw bank transactions could tell us more than quarterly financial statements ever could. So we built specialized LLMs trained on bank transaction data. With AI, we now construct credit-grade financials that are: ➡️granular (transaction-level) ➡️fresh (refresh daily) ➡️instantly verifiable (cannot be falsified) Then we layered on real-time signals from customer reviews and employee headcount changes that let us detect critical business shifts weeks or months before official reports. Our model dramatically cuts risk and cost. It opens up entirely new lending markets, segments previously labeled "too risky" or "not worth it." And this isn’t theoretical. Our models are assisting banks in underwriting millions of $ to real businesses, today. It reminds me of cloud computing replacing on prem services — a structural economic change, rather than a marginal improvement. If you're exploring similar shifts, reach out — I'd love to compare notes.

AI agents are having their moment in the spotlight—and in insurance, that’s raising an important question: What does agentic AI actually mean for underwriting? The answer isn’t replacement. It’s reallocation. Agentic AI isn't about removing underwriters from the equation. They’re about making sure the hours spent in a day better reflect where underwriting expertise is most valuable. Less time on manual, scattered processes. More time on decision-making and producer relationships. Where can Agentic AI play a role? Triage is a clear example. Today, many underwriters just work their email inbox top to bottom - because doing triage right is very hard. Figuring out which submissions deserve attention has always involved a mix of judgment and logistics. You need to collect data from multiple sources, brush the sub up against appetite, check for submission completeness, consider how a submission affects your portfolio, and determine how likely you are to bind the risk. In the past, if you wanted to triage submissions, a person had to do it. AI Agents can do all of this automatically. By centralizing fragmented data sources, applying context, and helping underwriters focus on the right opportunities, these systems improve the speed and precision of early decision-making—without taking away underwriter control. That’s the real promise of agentic AI in underwriting. Taking on critical, complex, time-consuming parts of the job, so that underwriters can focus on making better, faster underwriting decisions. #ai #insurance #underwriting