10.9.26 Economic Strength; Floify’s Maggie Swanson on Customizable Loan Apps; Prepayment Data

Robbie ChrismanWelcome to the Chrisman Commentary, Daily Mortgage News Podcast. I'm your host, Robbie Chrisman. Topics on today's episode include some interesting economic stats, who's prepaying what, and my interview with Floify's Maggie Swanson on creating fully customizable loan applications for any loan type, including HELOCs, construction, agricultural lending, non-QM, and more without custom development. Here, take a listen to a little preview. Robbie ChrismanWhat do you want people to understand about the cutting edge of tech some the stuff that Floify is working on, just where the mortgage industry is in terms of these uh gains that we're seeing? Maggie SwansonAI is a fun buzzword. We just want to make sure that at Floify we build it out methodically. We want to make sure that we make the lenders' lives easier rather than adding complication components that they have to deal with on a daily basis. So we want to make sure that it not only makes the borrower experience easier, we want to make sure AI makes the back end process easier by not adding additional compliance headaches or having to retract what we put out to make sure we do meet those compliance standards as they come up. Robbie ChrismanThanks to today's podcast sponsor, Floify, the mortgage industry's leading point of sale platform. Dynamic apps, which can be seen at Booth 600 during MBA annual next week, lets lenders create fully customizable loan applications for any type, including HELOCs, construction, agricultural lending, non-QM, and more without custom development. To learn more, visit Floify.com. We shouldn't be afraid to talk about the economy or politics, and there are plenty of clever sayings about economics. One of them is the stock market is not the economy. But a lot of us are wondering just how good is the U.S. economy? You wouldn't know it by stocks, and I read something from an economist that said U.S. equity market performance is increasingly being led by the magnificent seven, including hyperscalers, that's Amazon, Alphabet, Meta, and Microsoft, plus Apple, Nvidia, and Tesla. They closed yesterday with a combined market cap of almost $25 trillion, or 77% of US GDP. Moreover, year over year, 2026 third quarter tech EPS growth is expected to be 65%, double the overall S&P 500. The market may be a one-trick pony, but it's staggeringly large and profitable. According to Curinos Proprietary Application Index, September 2026, funded mortgage volume decreased 15% year over year and decreased 10% month over month. In the retail channel, funded volume decreased 15% year over year and decreased 7% month over month. The average 30-year conforming retail funded rate in September was 6.58, 8 basis points higher than August 2026, and 10 basis points higher than the same month last year. Purchase rates were 7 basis points higher month over month and 10 basis points higher year over year, while refinance rates were 15 basis points higher month over month and 20 basis points higher year over year. Curinos sources a statistically significant data set directly from lenders to produce these benchmark figures. U.S. Treasuries reversed early losses Thursday to finish and rally territory across the yield curve, despite pressure from rising oil prices, geopolitical concerns around Iran, and continued weakness in global sovereign markets. The turnaround accelerated after President Trump said there were no plans for additional attacks on Iran ahead of the midterms, easing fears of further escalation. Initial jobs claims fell to 197,000 for the week ending October 3rd, below expectations. Layoffs remain historically low, and labor demand continues to hold up despite some softening and continuing claims. Wholesale inventories rose 0.5% in August, below expectations, but following a revised 1.4% increase in July, suggesting inventory accumulation remains elevated but is moderating. Strong demand at yesterday's 30-year treasury sale followed Wednesday's solid 10-year reopening, and the Treasury's $6 billion long-end buyback provided additional support, pushing longer maturities prices higher. The auction was solid with a 2.54X bid to cover, a 0.1 basis point stop through, and 93.2% of awards going to non-dealers, well above the 90% six auction average, reflecting strong demand, particularly from indirect bidders. Treasuries rallied into the auction as oil prices fell on reduced Iran war concerns, but the market gave back some of those gains afterwards as bonds cheapened in the follow-through. For today's interview, I wanted to welcome to the show Floify's Maggie Swanson to talk about creating fully customizable loan applications for any loan type, including HELOCs, construction, agricultural lending, non-QM, and more without custom development. She's enterprise account executive at Floify, having joined the company more than five years ago, and playing a crucial role in the successful launch of Floify Lender Education. She leads large enterprise accounts and sales training sessions, strengthening the company's market presence and reputation. Robbie ChrismanPeople that have been listening to this podcast all week have heard me talk about dynamic apps, which lets lenders create fully customizable loan applications for any loan type, including HELOCs, construction, agricultural lending, non-QM, and more. But Floify is also embedding AI into the point of sale. How do those two strategies work together? Maggie SwansonYeah, so they're complementary. Dynamic apps in terms of the right application structure flows up for the right loan purpose, such as HELOC, construction, agricultural lending, non-QM, et cetera. AI then operates behind the scenes to make that tailored workflow faster and cleaner. So as a borrower moves through a loan system, AI auto-fills the 1003 with borrower data, automates document cleanup, validates uploads, calculates income and DTI, and streamlined steps that traditionally slows teams down. Lenders are not just capturing the right information up front, they are processing it faster and more accurately. The result is they configurable front end powered by intelligent automation underneath. So we have been intentional about where AI operates. We're not putting a chatbot in front of borrowers in a heavily regulated environment. Instead, we focus AI on operational automation, including extracting and structuring data, auto-completing key portions of the 1003, and accelerating underwriting prep without introducing regulatory or fair lending uncertainty. Robbie ChrismanSo why embed AI directly into the point of sale instead of layering it on top of the loan origination system? Maggie SwansonBecause the POS is where the borrower experience begins. If AI sits on top of a downstream system, you are often adding another portal, another login, or another handoff. By embedding AI directly into the POS, Floify keeps the experience seamless. Borrowers upload documents and answer relevant questions while AI handles repetitive tasks in the background, such as organizing files, validating data, autofilling 1003 information, and supporting underwriting prep. We have also intentionally stayed away from deploying borrower-facing chatbots. In a compliance-heavy industry like mortgage, open-ended AI conversations can create disclosure, documentation, and fair lending concerns. Rather than introducing that uncertainty, we apply AI in controlled, measurable ways that improve file quality and speed while keeping lenders firmly in control of the borrower experience. It enhances the lender-borrower relationship rather than inserting an unpredictable AI layer between them. Robbie ChrismanSo I speak with a lot of people on this podcast, and obviously we're in the age of AI. There's a lot of buzzwords being thrown around out there. What makes Floify's AI approach different from the buzzword heavy tools we see elsewhere? Maggie SwansonFloify is not putting a flashy chatbot in front of borrowers and calling it innovation. As I've mentioned, the AI is built into a proven POS platform and focused on measurable operational gains. It adapts workflows based on loan purpose, automates manual steps like 1003 completion and income calculators, and strengthens compliance through built-in safeguards, including automatic demographic suppression when required. And because it is configurable without code, lenders can adjust workflows without waiting on engineering resources. While we believe AI and mortgage should reduce risk, it does not create new gray areas. This is why our approach prioritizes automation, accuracy, and confidence and compliance over headline grabbing chat-bot features. This is not AI for show. It is AI for reducing friction where it actually matters. Robbie ChrismanSo, how do dynamic apps and AI together impact efficiency and compliance? And I think that's where the industry is headed. It obviously went from here's what AI can do to here's what AI is doing. And now the next step is how we can really see efficiency gains, or as I'm asking in this question, compliance gains as well. Maggie SwansonYeah, dynamic apps ensures the right disclosures sections and questions appear for each loan type. AI then supports the file behind the scenes by automatically completing 1003 questions so borrowers do not have to, and then validating documents, organizing uploads, and accelerating underwriting prep. Operationally, this means fewer missing conditions and less rework. From a compliance standpoint, it means loan purpose-specific configurations paired with automated safeguards and without the added regulatory uncertainty that open-ended borrower AI conversations can introduce. Instead of adding complexity, the technology reduces risk while improving speed. Robbie ChrismanFinally, what should lenders realistically expect from combining configurable applications with embedded AI? Maggie SwansonLenders should expect cleaner files earlier in the process, faster borrower completion times, and fewer manual touch points for lending teams. Borrowers move application flows while AI finally autofills 1003 data and handles the repetitive document management tasks that used to consume hours of staff time, which is very precious. Dynamic apps make the application smarter by structure. So embedded AI speeds up the execution. So it's dynamic apps with the AI component to it. Together, they turn the POS into a scalable growth platform rather than just an intake form itself. Lenders can see both dynamic apps and Floify's embedded AI capabilities in action next week at MBA Annual, October 11th through the 14th at Booth 600. I'll say hello if I see a love easy. Robbie ChrismanSeptember mortgage prepayments continue to slow sharply, with Fannie Mae speeds's falling 7% month over month to 6.7 CPR, the slowest since April 2025, and Ginnie Mae Speeds declining 10% to 7.6 CPR as the mortgage or as mortgage rates climbed to 7.28% in October, from 6.43% in early July and 5.98% at the February low. With only about 0.5% of mortgages currently estimated to be economically attractive to refinance, prepayments are likely to decline further as year-end seasonality approaches. However, a meaningful drop in rates could quickly expand the refinance opportunity with roughly 29% of mortgages potentially in the money if rates fell to percentage points. My advice? Servicing heavy originators such as Rocket and Rithm offer a relatively attractive risk reward given the value of their mortgage servicing portfolios in a higher rate environment. With rates still biased higher and limited near-term catalysts for a reversal, prepayment speeds are likely to remain subdued, supporting mortgage servicing valuations but reducing refinancing activity. The latest prepayment data shows clear and persistent differences in servicer behavior. Rocket slash Quicken stands out as the consistently fastest payer among both UMBS 30 and 15-year loans, while Freedom and Fifth Third also rank among the fastest in 30-year mortgages. At the other end, Bank of America, loanDepot, and Provident are frequently among the slowest 30-year servicers, while Lakeview, JP Morgan, Chase, and Provident lag in 15-year loans. By loan age, American and Rocket Quicken lead the early stage 30-year universe with the fastest speeds concentrated in the 24-36 WALA buckets. Conversely, Idaho HFA has remained among the slowest for 14 consecutive months, highlighting meaningful and persistent servicer level differences in prepayment speeds. Today sees an extremely light economic calendar with preliminary University of Michigan consumer sentiment the main release. Boston Fed President Collins will also deliver remarks. We began Friday with agency MBS prices slightly worse than yesterday's close, the two-year yielding 4.79%, and the 10-year yielding 5.25% after closing yesterday at 5.23%. Let's wrap up with a joke and some housekeeping. Do you hear in Chicago the local police are hunting the knitting needle nutcase, who has stabbed six people in the rump in the last 48 hours? They believe the attacker could be following some sort of pattern. Thanks again to Floify for sponsoring this week's podcasts. Floify is the mortgage industry's leading point of sale platform and dynamic apps, which can be seen at Booth 600 during MBA annual next week. Let lenders create fully customizable loan applications for any loan type, including HELOCs, construction, agricultural lending, non-QM, and more without custom development. To learn more, visit Floify.com.