 |
The Mortgage AI PodcastAuthor: Frazier
The Mortgage AI Podcast is a show for lenders, mortgage brokers, loan officers, and operators who want to leverage AI for growth instead of fearing it. Join Jason Frazier for episodes that dive into the tools, strategies, and honest conversations about AI implementation in the mortgage space. We cut through the hype and nonsense to focus on what's actually working today and what you need to know tomorrow.The Mortgage AI Podcast is presented by DIFRNT Coaching and is sponsored by MortgageCon Language: en-us Genres: Business, Entrepreneurship, Marketing Contact email: Get it Feed URL: Get it iTunes ID: Get it |
Listen Now...
From Microsoft and OpenAI to Rebuilding Mortgage | MAP Ep. 5
Episode 5
Thursday, 1 October, 2026
Send us Fan MailThe mortgage industry has a bad habit of taking new technology and forcing it to fit processes that should have been rebuilt years ago.AI can make that mess move faster. It can also give us an opportunity to rethink how mortgage origination should work from the ground up.In this episode of The Mortgage AI Podcast, I sit down with Oren Michaely, CEO and co-founder of Elio Mortgage.Oren worked at Microsoft Research and helped with the Microsoft and OpenAI collaboration before ChatGPT was released to the public. He later worked at Motive Partners, where his job was to help financial services companies use AI to create measurable business value.Then he left the comfortable corporate world and entered the beautifully complicated mortgage industry.Why mortgage?Oren saw an industry that has spent decades trying to scale by adding more people, more systems, and more layers. Loan officers work across lender portals, CRMs, LOS platforms, pricing engines, email, text messages, phone calls, and spreadsheets.Every handoff creates another chance for information to disappear or a mistake to happen.Elio is building an agentic mortgage company where AI operates as an invisible layer beneath the people. The loan officer becomes the air traffic controller of the transaction while the platform handles much of the repetitive work required to move the file forward.That includes dynamic borrower applications, automatic document requests, document classification, income calculations, pre-filled applications, pricing, guideline searches, pre-approvals, and communication throughout the loan process.A DSCR borrower, for example, should never waste time answering questions about personal income that have nothing to do with the loan. The application should understand the scenario and request only the information needed to move forward.The same intelligence can help loan officers manage their entire book of business. Elio’s platform is being built to identify refinance opportunities, improve outreach to Clients and referral partners, and use data to help loan officers decide who to contact and why.I called it portfolio intelligence during our conversation. Oren and his team are building a workspace that understands the loan officer’s business and helps them manage it without stitching together 14 different tools.We also break down how to build an AI agent without creating a digital employee that has no idea what the hell it is doing.Oren recommends breaking a process into small units of work.Give the AI one specific task. Validate the result. Keep a human involved while you refine it. Then connect that task to the next step in the process.Telling ChatGPT to “build me a transaction coordinator” skips every important detail the system needs to perform the job correctly.In this episode, we cover:Oren’s work with Microsoft and OpenAI before ChatGPTWhy he chose the mortgage industryWhat an agentic mortgage company looks likeWhy outdated processes limit the value of new technologyHow to break workflows into units of workThe importance of keeping humans involved during developmentWhy data remains the biggest obstacle to reliable mortgage AIHow Elio is rebuilding the LOS around an AI-ready data structureDynamic applications and automated document collectionUsing AI to pre-calculate income and prepare files for reviewGiving loan officers intelligence across their entire portfolioWhy complex vibe-coded backends can create serious security risksWhat a headless mortgage platform could look likeHow loan officers can begin experimenting with AI this weekOren’s advice for getting started is simple: spend two or three hours working with the technology.Tell the model what you do. Explain your role. Ask it where you may be wasting time or creating unnecessary work. Choose one process and build something small enough to test.You do not need an engineering team to begin understanding what is possible. You need curiosity, a clear goal, and enough time with the tool to move beyond the basic prompt window.00:00 Podcast Welcome01:19 Meet Oren01:38 AI Career Journey06:45 Why Mortgages09:46 Building Agentic Workflows16:06 Data Context Challenge20:20 Elio Platform Vision22:02 Automated Doc Intake22:33 Instant Preapproval Workflow23:07 Loan Officer Leverage24:03 Refi Radar Integrations24:46 Pricing Guidelines Engine25:34 Portfolio Intelligence Platform27:35 Stealth Launch Iteration30:06 AI Hub Future UI34:37 Security Compliance Focus35:05 Rapid Fire AI Takes37:17 Hands On AI Advice38:42 Closing Thanks OutroDiscover Elio MortgagePresented by: DIFRNT CoachingFounding Sponsor: MortgageConBroker Sponsor: Summit LendingFriends of the Program: Empower LO
