Beyond the Hype: Real-World AI Adoption in the Mortgage Industry

Esteemed industry speakers get ready for IMN's debuting Mortgage AI conference by sharing early answers to pressing questions in this rapidly evolving space. Join industry professionals on October 9-10 at the Waldorf Astoria Monarch Beach, for further strategy informing insights - register for Mortgage AI today!

Nik Shah, CEO, 100x

1. How are you using AI? We’re replacing human bottlenecks with AI. Our agents — like Sophie in origination, Sally in home sales, and Nikolai as our AI CEO — run end-to-end workflows, not just copilots but full operators.

Q: Biggest impediments to adapting AI? The hardest part isn’t tech, it’s trust. We won trust by proving AI could deliver 100x outcomes.

Q: Build, buy, or hybrid? Hybrid. We leverage frontier models but wrap them in our own Small Language Models (SLMs), trained on real estate and sales data.

Q: Challenges that caught you off guard? Human workflows are messy, so our AI isn’t just logical but also empathetic, capable of guiding emotional, high-stakes decisions in complex conversations.

Q: Keeping non-technical stakeholders engaged? Don't buy AI, instead buy revenue.

Q: How do you measure success? Same as a human: revenue & cost.

Q: Advice for starting the AI journey? If your CEO or leader doesn't use AI, you will (most likely) fail. AI isn't a tool, it's a culture shift.

Q: Highlight of your summer? Watching Sasha, our AI Sales Bot, helping seasoned executives realize: this is the future.

Neena Vlamis, President, A & N Mortgage Services Inc.
1. How are you using AI? AI is integrated into daily workflows for tasks such as research, document review, contract analysis, compliance monitoring, and risk assessment. It automates repetitive processes, enhances decision-making with predictive analytics, and streamlines communication by generating summaries and drafting documents. This allows for greater efficiency and accuracy in handling complex legal and regulatory matters.

2. Largest Impediments to Adapting AI & Overcoming Them?

Impediments:

• Data Quality & Availability: Ensuring access to clean, relevant, and secure data is often a major hurdle.
• Change Management: Resistance from staff accustomed to traditional methods.
• Integration with Legacy Systems: Technical challenges in connecting AI tools with existing infrastructure.
• Regulatory & Ethical Concerns: Navigating compliance, privacy, and ethical use of AI.

Overcoming Strategies:

• Invested in robust data governance and security protocols.
• Provided comprehensive training and clear communication about AI’s benefits.
• Adopted modular AI solutions that integrate smoothly with legacy systems.
• Working to establish an internal ethics committee to oversee responsible AI use
  

3. Build, Buy, or Hybrid Approach? A hybrid approach is most effective. Core AI capabilities were built in-house to ensure customization and control, while specialized tools (e.g., for natural language processing or document automation) were acquired from trusted vendors. This balances innovation, speed, and reliability.

4. Unexpected Technical Challenges? Model Interpretability: Technical teams found it challenging to explain AI decisions to non-tech people.

• Data Labeling: The volume and complexity of documents made accurate data labeling more difficult than anticipated.
• Scalability: Scaling AI solutions to handle large, diverse datasets required more infrastructure investment than initially planned.

5. Engaging Non-Technical People - working on this:

• There is always one! Regular workshops and demonstrations showing AI’s practical impact.
• Transparent communication about project goals, risks, and progress.
• Involving discussions in pilot projects and feedback sessions.
• Sharing success stories and metrics that matter to their roles (e.g., time saved, error reduction).

6. Success is measured by:

• Efficiency Gains: Reduction in time spent on manual tasks.
• Accuracy: Improved quality and consistency of outputs.
• User Adoption: Engagement and satisfaction among staff.
• Business Impact: Tangible improvements in compliance, risk mitigation, and client outcomes.

7. Advice for Starting an AI Journey:

• Start Small: Pilot AI on a manageable, high-impact use case.
• Focus on Data: Invest early in data quality and governance.
• Engage Team: Build cross-functional teams and communicate benefits clearly.
• Iterate & Learn: Expect setbacks, learn from them, and refine your approach.
• Stay Ethical: Prioritize transparency, fairness, and compliance from the outset.

8. Highlight of the Summer? The highlight was leading a cross-functional workshop where we went to Texas and business teams collaborated to design an AI-powered compliance dashboard. It’s a work in progress!

Abby Shemesh, Chief Acquisitions Officer, AMERIN note xchange

1. Can you let me know about how you are using AI? Customer service, chatbots, research.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? The biggest impediments have been with accuracy of Large Language Model AI (LLMs) and trusting the information. In order to overcome, we need to take extra steps by human verification.

3. Did you build, buy or use some kind of hybrid approach to build your system? Both - we used an AI engineer and purchased white label.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Accuracy in the info when using for research and diligence.

5. How did you keep non-technical stakeholders engaged in the AI journey? By increasing productivity while decreasing costs and overhead.

6. How do you measure success? Reduction of costs and augmentation of abilities beyond the competition.

7. What advice would you give to someone just starting their AI journey? It is all about the prompts you give the AI - BE VERY SPECIFIC.

8. What was the highlight of your summer? Beach time!

Alysse Prosnick, EVP of Operations, Angel Oak

1. Can you let me know about how you are using AI? We're leveraging AI to streamline and enhance several operational workflows. Most notably, we've implemented AI-powered systems to replace legacy tools. These new systems improve speed and accuracy in our processes. We're also using AI to support monthly KPI analysis, ensuring consistent formatting and insights across all operational stages.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? One of the biggest challenges has been change management—getting teams comfortable with new tools and workflows. We addressed this by involving stakeholders early, offering hands-on training, and demonstrating clear value through pilot programs. Another hurdle was integrating AI into legacy systems, which required thoughtful architecture and collaboration between technical and operational teams.

3. Did you build, buy or use some kind of hybrid approach to build your system? We took a hybrid approach. We built custom components tailored to our operational needs while integrating third-party AI platforms where it made sense. This allowed us to maintain control over critical workflows while accelerating deployment timelines.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Yes—data normalization across systems proved more complex than anticipated. Our technical team had to navigate inconsistencies in data formats and legacy logic that weren’t well documented. It was a good reminder of the importance of clean data and cross-functional collaboration.

5. How did you keep non-technical stakeholders engaged in the AI journey? We focused on transparency and relevance. Regular updates, demos, and real-world use cases helped bridge the gap. We also made sure to tie AI initiatives directly to business outcomes—like faster disclosures, more polished responses or improved accuracy—so stakeholders could see the impact firsthand.

6. How do you measure success? Success is measured through a combination of operational KPIs and user adoption. For example, we track improvements in turnaround time, error rates, and efficiency. We also monitor engagement metrics to ensure teams are actually using and benefiting from the AI tools.

7. What advice would you give to someone just starting their AI journey? Start with a clear problem to solve and build from there. Don’t chase AI for the sake of it—focus on where it can add real value. Also, invest in team awareness and engagement early. The tech is only part of the equation; adoption is what drives impact.

8. What was the highlight of your summer? Professionally, seeing our new AI systems go live and watching the team embrace them was incredibly rewarding. It was also delightful to see how the team adapted with the increased time savings in their day from mundane task-based work to a more meaningful customer experience focus.

Ashlei McAleer, Chief of Staff, Angel Oak

1. Can you let me know about how you are using AI? Absolutely. We’re using AI to streamline decision-making, improve operational efficiency, and enhance strategic planning. It’s especially helpful in surfacing insights from large datasets, automating repetitive tasks, and supporting cross-functional collaboration.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? One of the biggest challenges has been change management—getting teams comfortable with AI tools and trusting their outputs. We’ve overcome this by focusing on transparency and ensuring that AI augments rather than replaces human judgment. Understanding what is available and how best to select AI products has also been challenging. We also established a Techology Leadership Committee to help evaluate technology risk and investments, foster collaboration and align goals with the business needs.

3. Did you build, buy or use some kind of hybrid approach to build your system? We’ve taken a hybrid approach. We leverage existing platforms for foundational capabilities, but we’ve also built custom layers to tailor AI to our specific workflows and data needs. This gives us flexibility while still benefiting from proven infrastructure.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Yes—data quality and integration were more complex than anticipated. Even with strong infrastructure, aligning disparate data sources and ensuring consistency across systems required more effort than expected. It was a good reminder that AI is only as good as the data it’s built on.

5. How did you keep non-technical stakeholders engaged in the AI journey? We made it real for them. Instead of talking about models and algorithms, we focused on outcomes—how AI could save time, reduce errors, or unlock new opportunities. We also involved them early in pilot programs so they could see the impact firsthand and help shape the direction.

6. How do you measure success? Success is measured by adoption, impact, and alignment with strategic goals. We look at usage metrics, time saved, decision quality, and stakeholder feedback. But we also measure success by how AI helps us stay agile and forward-looking in a rapidly changing environment.

7. What advice would you give to someone just starting their AI journey? Start with a clear problem to solve. Don’t chase AI for the sake of it—focus on where it can add real value. Build cross-functional teams, invest in data readiness, and don’t underestimate the importance of change management. And stay curious—this space evolves fast.

8. What was the highlight of your summer? Personally, spending time with family and recharging outdoors has been rewarding.

Murat Gurer, Director, Best In Funding, Inc.

1. What advice would you give to someone just starting their AI journey? AI is a vast and exciting field, but it’s easy to get overwhelmed. My advice? Start by understanding your own goals. Not all AI tools or solutions fit everyone—each person’s needs, interests, and learning style are unique. So:

• Identify your focus area
• Be intentional: Don’t chase every shiny tool. Find what aligns with your goals and go deep.
• Learn by doing: Tutorials and courses are great, but building small projects will teach you more than theory ever could
• Stay curious and connected: Join communities, follow thought leaders, and keep up with trends. AI evolves fast.
• Efficiency matters: Use tools that streamline your workflow and help you learn faster—whether that’s coding platforms, visualization tools, or AI companions like copilot (my favorite) ,chatgpt, gemini..etc

What was the highlight of your summer? Week Eurotrip and a family reunion!

Chad Smith, President & Chief Operating Officer, Better Mortgage

1. How is Better.com using AI? At Better.com, we leverage AI through our proprietary Tinman® platform, which is an loan origination engine that automates underwriting, document processing, income verification, and more. This platform allows our loan consultants to be over three times more productive than the industry median while reducing fulfillment costs by about 35%. In addition, we recently launched Betsy™, our AI loan assistant, which is integrated into Tinman. Betsy can converse naturally with customers, provide real-time updates, collect information, process HELOC approvals, generate rate quotes, and even lock rates. As of this year, Betsy is handling over 127,000 customer interactions each month, and about 40% of our loan files are reviewed through automated systems.

2. What have been some of the largest impediments to adapting AI, and how have you overcome them? One of the largest impediments businesses like ours face is the complexity of integrating AI across multiple systems such as CRM, pricing engines, document management, and point-of-sale software. We overcame this by consolidating everything into Tinman, which serves as our unified AI backbone. Another challenge was addressing trust, privacy, and fairness in mortgage lending, since we work with sensitive financial data under heavy regulatory oversight. To overcome this, we invested in strong encryption, centralized data management, and transparency across our processes. Finally, we learned early on that customers found simple chatbots frustrating, so we pivoted toward building Betsy, a conversational, voice-based assistant that delivers meaningful, human-like interactions.

3. Did you build, buy, or use a hybrid approach to develop your AI systems? Our AI platform is built in-house. Both our proprietary platform, Tinman, and Betsy were developed by our internal teams, which allowed us to maintain control of the technology and tailor it to the unique challenges of mortgage lending. By building in house, we’ve been able to build our tech to meet the needs of our loan officers. Building internally also gave us the flexibility to adapt quickly as customer needs and regulatory requirements evolved.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? One of the more difficult challenges we encountered was scaling a voice-based AI assistant in such a highly regulated and high-stakes industry. When Betsy began handling critical tasks like approving HELOCs and locking rates, the technology had to meet extremely high standards for accuracy and compliance. Our team quickly realized that it was not enough to simply automate processes, we had to ensure that the AI provided empathy and trust in every interaction.

5. How did you keep non-technical stakeholders engaged in the AI journey? We kept non-technical stakeholders engaged by consistently communicating the measurable impact of AI across the business. For example, we showed how productivity had tripled and how fulfillment costs had dropped by more than a third. We also highlighted external recognition, such as Betsy being named a finalist for the 2025 Inman Innovator Awards, which validated our progress and energized the organization. Most importantly, we framed AI not as a replacement for people but as a tool that empowers our employees by removing repetitive tasks and allowing them to focus on customer relationships.

6. How do you measure success? We measure success through both quantitative and qualitative outcomes. Quantitatively, we look at metrics such as productivity, fulfillment cost reductions, automation rates, and customer interactions. For instance, Betsy now manages more than 115,000 monthly interactions, and Tinman automates 40% of loan reviews. Qualitatively, we measure success by customer experience improvements, like being able to deliver a “One Day Mortgage” approval. We also take pride in recognition from the industry, such as our Innovator Award nomination, which reflects that we are moving the entire mortgage space forward.

7. What advice would you give to someone just starting their AI journey? My advice would be to focus AI on mission-critical workflows where it can create the most impact. At Better.com, that initially meant using AI to streamline underwriting and customer engagement rather than trying to apply it everywhere at once. I would also stress that AI should be designed to empower employees, not replace them, by automating repetitive tasks so people can focus on the human side of the business. Measuring clear ROI is crucial, whether in productivity gains, cost savings, or customer satisfaction. Finally, I would emphasize the importance of building trust, ensuring data security, and celebrating milestones along the way to keep your teams and stakeholders engaged.

8. What was the highlight of your summer? The highlight of our summer was Better.com being named a finalist in the “Most Innovative Use of AI” category at the 2025 Inman Innovator Awards. That recognition was incredibly rewarding for our teams, because it validated both the technical breakthroughs we have made and the way we are using AI to improve the customer experience. We were also proud to report 166% YOY growth on our home equity products in our Q2 earnings, with $80M a month in HELOC and HELOAN originations, fueled by our 1 Day HELOC product.

Keith Soura, V.P. Engineering, Better Mortgage

1. How are you using AI? At Better, we're pioneering a human-centric AI model that amplifies both customer and employee experiences. Our agent Betsy is a prime example—seamlessly delivering voice and text support that ranges from rate inquiries to intricate financial modeling, while supporting staff with fraud detection, scheduling, and code reviews. This aligns with industry shifts toward agentic systems that enhance productivity across operations, HR, customer service, and finance. But unlike one-size-fits-all narratives, we view AI as a precise toolbelt, intelligently applied to maximize value and drive focused innovation.

2. What have been the biggest impediments to adopting AI, and how did you overcome them?

Cost & Compliance: AI investments remain substantial and highly scrutinized in regulated sectors. Real-world implementations often falter without strong legal frameworks and disciplined ROI planning. We counteract this with strategic procurement, legal collaboration, and rigorous ROI trails.

Adoption & Culture: Across financial services, cultural resistance and fragmented data slow progress. We address this by democratizing access and gamifying adoption, fueling an internal culture shift that rewards creativity and positive impact on our KPIs.

Governance & Risk: Regulatory complexity and the sheer scale of AI can threaten operational trust. Experts underscore the importance of embedding governance frameworks that manage risk without throttling agility. We solve this by appointing a dedicated special projects manager to ensure visibility and compliance while empowering innovation.

3. Did you build, buy, or take a hybrid approach? We’ve adopted a pragmatic hybrid model. Whenever a capability amplifies our IP or core competencies, we build; otherwise, we buy. This enables us to create a unified, seamless backend experience while capitalizing on best-of-breed off-the-shelf tools. It’s a balanced strategy that aligns with the strategic notion of using AI both as a platform and capability enabler.

4. Were there any technical challenges that caught your teams off guard? The biggest surprise? The variance in model performance—even among leading edge models—was stark. This reflects broader industry observations around foundation models and their unpredictable behavior in deployment environments. To manage this, we’ve implemented a holistic evaluation framework—closing the gap between experimentation and production stability while maintaining agility.

5. How did you keep non-technical stakeholders engaged? By embedding democratization and gamification into the rollout. We broke down silos, encouraged cross-department collaboration, and highlighted innovations from every corner of the company. This mirrors industry best practices that show engaged stakeholders and transparency are critical for scaling AI responsibly. Slack channels, newsletters, and all-hands celebrations transformed AI from an abstract concept into a shared mission.

6. How do you measure success? Our North Star metrics are rooted in business fundamentals: origination volume, lending margin, and cost per loan. We shed the “shiny object” syndrome by ensuring every AI project directly aligns with these KPIs. Benchmarks show that organizations with clear objectives see ~15.8% revenue uplift, ~15.2% cost reductions, and ~22.6% productivity gains—and these gains emerge when AI solves specific, measurable problems – looking at our quarterly statements we are exceeding even these expectations with our AI program. We also track employee-level efficiency gains to ensure that productivity improvements translate into consumer value and operational scalability.

7. What advice would you give to someone starting their AI journey? Be realistic and empowering. AI shouldn’t be portrayed as replacing humans—it’s about enabling them. Frame it as a force multiplier: an opportunity to supercharge—not sideline—your teams. Build a human-first culture that encourages experimentation and learning. In line with expert advice, view AI agents as ways to "elevate the worker" rather than replace them. Create energy, not fear—because the real narrative isn’t about AI taking over; it’s about AI lifting everyone forward.

8. What was the highlight of your summer? It was a lovely summer! Being outside, enjoying New York City (my home), and taking advantage of all of the great summertime art programs available in parks around NYC with my family was my highlight. I was also able to celebrate a big birthday – 40 this year – with my closest friends and family in Abu Dhabi and that was a lifetime memory for me that I’ll cherish forever!

Damon Germanides, Co-Founder & Broker, Insignia Mortgage

1. Can you let me know about how you are using AI? As a boutique broker and lender, I have been focused on brining AI tools into our organization very carefully. Some of the off-the-shelf products we utilize such as Salesforce and Office 365 already have AI embedded into their systems, but as an organization we are starting to realize that there are many repetitive tasks that can be streamlined with the use of AI. Examples include reviewing an operating agreement or trust for consistent signature lines or ownership interest or scanning title reports for tax liens. We recently created a test case where we loaded up all of our lending guidelines into a GPT, gave the GPT very explicit instructions and data and will be testing that tool with our underwriting team and loan officer team in the fall.

From a sales standpoint, bringing in AI Agents to help our loan brokers book appointments, respond more quickly to inquiries and provide a better borrower experience is also top of mind for our organization. While this is a work in progress, utilizing AI agents has tremendous potential to make our sales team more efficient and better and more timely with follow up.

Finally, in order to implement AI, you have to understand how it works. This summer I spent two days in an AI intensive designed around mortgage. This really helped me better understand the tools. While our whole team cannot attend these in person events, I have shared some very good youtube videos that break down use cases for AI so the team also gets comfortable around bringing AI into the organization.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? The loan origination business is not a young person’s business at the moment. I believe I read a stat not too long ago that the average loan officer is over 50 years old. Bringing in highly technical and disruptive technology into an organization with a middle-aged work force takes patience, education and a well thought out process that does not overwhelm both the underwriting team and sales team. My AI plan is akin to “chopping wood” as we want to focus on doing things log by log and not get overwhelmed by trying to do too much all at once.

Dame Jen Du Plessis, CEO, Kinetic Spark Consulting

1. Can you let me know about how you are using AI? AI isn’t here to replace us—it’s here to release us. In my business, I use AI to automate low-level task such as scheduling, client follow-up, and marketing of not only used to eat up hours my team, but my team’s as well. That gives me freedom to focus on relationships, coaching, and creating strategic partnerships—the activities that drive revenue and legacy. AI helps me scale without sacrificing humanity, and that’s the balance most business owners are desperate to find.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? The biggest impediment is fear of the unknown or something new. I hear all the time from business owners two things:

a. Ai will take jobs away from us as humans, and
b. I don’t like Ai because I don’t understand it. I overcame that by reframing it: AI doesn’t replace your value, it amplifies it. Once my team saw AI could do the repetitive, draining tasks, they leaned into it. I remind leaders that people don’t fear the tool, they fear losing their identity. Your job is to show them AI secures their role by elevating what only humans can do.

3. Did you build, buy or use some kind of hybrid approach to build your system? I took a hybrid approach. I bought ready-made tools for efficiency but built custom AI workflows around my proven frameworks like DRIVE and Hustle to Harmoney. What I learned is that it’s important not to force yourself into someone else’s system. It’s akin to a CRM, don’t purchase one until you know what you are looking for it to achieve. Start with your genius, then let AI amplify it! This way, the tech fits your business rather than your business yielding to the tech.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Absolutely. We assumed the hardest part would be the tech. Turns out; the challenge was the translation—getting my vision as a coach and strategist into a system my technical team could code and automate. We had to learn to speak each other’s language. For leaders, never abdicate your vision to the tech team. Own the translation, or your genius never makes it into the Ai machine. Upload everything you’ve said, videos, interviews, books, articles, emails so Ai can learn you, your values, voice, and vision.

5. How did you keep non-technical stakeholders engaged in the AI journey? Storytelling. I painted a picture of what life would look like after AI—more time, more freedom, more money, less burnout. Every time we introduced a new tool, I tied it back to their personal benefit. If your processor hears 'automation' and thinks 'job loss,' you’ve lost them. But if they hear 'automation' and think 'I finally get home for dinner,' you’ve got lifelong buy-in. The biggest issue here for a business to thrive and not survive, is that early adoption is key. You must gain the edge in today’s market! The longer you wait, the harder it will be to adapt and the gap to success will be larger than ever before due to the pace of change.

6. How do you measure success? I measure success two ways: by numbers and by my lifestyle. Yes, I track efficiency, leads, revenue, efficiencies, and ROI. But the real measure? My clients and my team telling me they’re working fewer hours, closing more business, and finally taking that vacation they have put off for years. AI success isn’t just productivity—it’s profitability with freedom to enjoy what we are all building.

7. What advice would you give to someone just starting their AI journey? Start small and start human. Pick one area where AI removes friction—maybe it’s your CRM follow-up/email campaigns, maybe it’s your marketing content/engaging social media posts that result in action and profit—and prove to yourself and your team it works. Then expand. And please remember: AI is not the driver of your business. You are the driver. AI is the power steering that makes the drive smoother. If you get that right, you’ll scale without stress and grow without burnout or sacrificing your persona life.

8. What was the highlight of your summer? I’d say what wasn’t! We spend a lot of time on our boat – the weather was great this summer. But the best highlight was spending 3 weeks in Hawaii with our children and grandchildren exploring and experiencing so many new things – challenging each of us to push beyond our comfort zones- while visiting 3 different islands. Wow! A vacation to remember!

Alec Hanson, SVP, Head of Revenue Development & Growth, loanDepot

1. Can you let me know about how you are using AI? Daily. Research, Marketing Data and content ideas, current events/news, competitor information, etc…

2. What have been some of the largest impediments of adapting AI? How have you overcome them? Legal and compliance concerns over risk and exposure. AI’s access to private customers information, ongoing issue. Not solved.

3. Did you build, buy or use some kind of hybrid approach to build your system? Both.

4. How did you keep non-technical stakeholders engaged in the AI journey? Showcase use cases, talk about effects and enhancements over technical requirements.

5. How do you measure success? Productivity pick ups, speed to decisions, turn time improvements, etc..

6. What advice would you give to someone just starting their AI journey? Start immediately, don’t wait.

7. What was the highlight of your summer? 20 year anniversary trip back to Greece where my wife and I spend our honeymoon.

Russell Petty, CTO, Milestone Mortgage

1. How do you measure success? I use Ai almost every moment of the day. Ai categorizes my calendar in the mornings so I know what to do that day. It filters out my inbox so I only see emails I need to respond to, which it has already drafted replies for me in my voice upon opening. I use it to code, to deploy software, and to build tools to increase productivity in my team.

2. What have been some of the largest impediments of adapting AI? How have you overcome them?
The largest impediments in adapting Ai have been getting teams to utilize the tools rather than falling back to what "used to work" and ensuring that any models are secured, locally hosted properly, and scalable. A lot can be done with Ai, but at times the security requirements and user buy in are the hardest parts.

3. Did you build, buy or use some kind of hybrid approach to build your system? Our entire system, from the POS to CRM to "lite-LOS" is all Ai based. We built all of it from scratch internally. Nearly every single piece of tech we use is built, not bought.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Ai coding, while it seems to be an amazing thing, actually held our team back a bit when the projects reached a certain level of complexity. I think it is great for getting a quick MVP up, but anything beyond that, for our team, showed a slow down in productivity.

5. How did you keep non-technical stakeholders engaged in the AI journey? To keep non-technical stakeholders engaged, it's important to really drive home how it effects **them**. What problem are we solving? They do not care about RAG, LangChain vs LangGraph, or how many billion parameters the model is. Sell the solution, not the architecture.

6. How do you measure success? I measure success by how many days in a row I can live my perfect day.

7. What advice would you give to someone just starting their AI journey? For someone just starting their Ai journey, stop reading and start doing. Break things. Watch youtube, read Reddits, make a HuggingFace account. Start with ChatGPT, learn custom GPTs, then move on to Agents with N8N, then start vibe coding with Cline, Cursor, or Claude Code. Learn how to deploy apps on services like Railway, Netlify, or Vercel. Learn backend, databases, and how they work. You do not need to know how to code to do Ai. You DO need to know security.

8. What was the highlight of your summer? Highlight of my summer was the launch of Broker Bot - we helped loan officers link with over 30,000 lenders in just 7 days - all using Ai.

Scott DiGregorio, Branch Manager & Mortgage Loan Officer, NEO Home Loans

1. Can you let me know about how you are using AI? AI has become a central part of the branch and the team’s workflow - creating pipeline meeting action steps for the team, using it as a sales manager to analyze team member’s client conversations, realtor searches with similar interests, marketing strategies, identifying clients DISC profiles and helping to customize communication in accordance with their DISC - and so much more

2. What have been some of the largest impediments of adapting AI? How have you overcome them? The biggest obstacle that I experienced and have seen in many others is fear and we just decided that we were going to work with it until that fear was replaced by excitement and that is exactly what happened.

3. Did you build, buy or use some kind of hybrid approach to build your system? Built it.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Sure, the first AI agent took some meddling with for sure but we persevered through it.

5. How did you keep non-technical stakeholders engaged in the AI journey? Getting them involved was the key - if we were doing it together, we all had a stake in it.

6. How do you measure success? Hours saved, per loan profitability growth.

7. What advice would you give to someone just starting their AI journey? Jump in with both feet.

8. What was the highlight of your summer? Getting an AI built that can run numbers for clients based on homes they were interested in. Personally - visiting my daughter who recently graduated UF and moved to Atlanta : )

Kevin Peranio, Chief Lending Officer, PRMG

1. Can you let me know about how you are using AI? For five years we have used an AI chat bot to automate all of our e-mail helpdesk ticketing systems. We're in the process of implementing AI voice technology. We also use some client facing technology as well. I have demoed dozens of AI technology pieces.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? The initial lift getting training data into the AI takes about a month but after that we start to yield immediate return on investment.

3. Did you build, buy or use some kind of hybrid approach to build your system? Buy.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? Driving adoption especially among originators is the biggest challenge.

5. How did you keep non-technical stakeholders engaged in the AI journey? Ongoing training and demonstration of the benefits helps a lot.

6. How do you measure success? Usage and cost savings.

7. What advice would you give to someone just starting their AI journey? Call me.

8. What was the highlight of your summer? AI Voice.

Debbie Michel, VP-Advanced Analytics and AI, UWM

1. Can you let me know about how you are using AI? At UWM, AI is at the center of everything that we do. We’ve built a generative AI platform that connects multiple LLMs with our core systems and are leveraging AI to transform the mortgage process by accelerating our Team Members in their day-to-day roles and unleashing powerful, new AI-driven tools for our broker partners.

2. What have been some of the largest impediments of adapting AI? How have you overcome them? A lack of clean data and well-defined workflows. This is still a work in progress. The first step is education and awareness to help all our teams, both technical and business, understand how AI works best and transform our processes, systems, and data to truly harness the power of these tools.

3. Did you build, buy or use some kind of hybrid approach to build your system? We have a hybrid approach.

4. Were there any concepts or challenges that your technical team found unexpectedly difficult to navigate? Caught them off guard? The most obvious is just that this is a net new technology. The biggest surprise, though, probably wasn’t the technology itself but the governance side of these technologies. The biggest riddle has been how to make AI outputs explainable, consistent, and compliant in an industry as regulated as mortgage. Balancing innovation with responsibility and speed with oversight has been difficult and forced us to think in new ways and adopt new frameworks.

5. How did you keep non-technical stakeholders engaged in the AI journey? We bring our stakeholders along for the journey from ideation to solutioning and then throughout execution and implementation. Our stakeholders are involved in working sessions to define the pain point and hone in on solutions to solve those pain points and metrics to measure the outcomes. Once we’re aligned on the solution, stakeholders are shown bi-weekly demos to highlight progress and stay close to the development process.

6. How do you measure success? I would say this depends on the work. Generally, though, we measure success by establishing objectives and key results for work that we have picked up.

7. What advice would you give to someone just starting their AI journey? Start small and with specific pain points and outcomes / success metrics in mind. Choose something with low complexity to help your tech teams better understand the technology and the capabilities. This is truly paradigm shifting; expect there will be a learning curve for your technologists and business partners. Celebrate wins and also pivots – knowing what this technology is NOT good for is just as important, if not more so, than knowing what it is good for.

8. What was the highlight of your summer? Releasing LEO (Loan Estimate Optimizer) to the broker community, which allows a broker to drag and drop a competitor’s Loan Estimate (LE) and then receive a line-by-line breakdown, This breakdown highlights misleading or incomplete information, flags gaps and opportunities that would allow the broker to offer a better deal (for example, by reducing title fees via TRAC or applying appraisal credits), and equips brokers with talking points for a borrower.

Join industry professionals on October 9-10 at the Waldorf Astoria Monarch Beach for further strategy informing insights - register for Mortgage AI today!