Back to Industry Trends & Insights

MReport Webinar – Industry Experts Discuss AI in Mortgages

Mreport Icon

Four Executives Weigh in on AI’s Challenges, Misconceptions and Momentum

Recently, MReport hosted a webinar with industry executives from the banking, lending and technology sectors to talk about AI in the mortgage industry today. They discussed a wide range of topics, including the challenges with AI in the mortgage business, what some of the common misconceptions are, and where we are gaining momentum with AI deployments. 

Here are some of the main topics that were discussed, with some quick highlights and excerpts for each topic. To hear the full detail, listen to the webinar from MReport.

The Challenges for AI in Mortgages

  • Accessibility to useful data is still an obstacle. There is a growing awareness of the power of data and how it can drive AI and automation.
  • We need to figure out how to build AI models on top of the data that we have.
  • First, look at the business problems that you need to solve, then develop processes and systems to fix that problem. For example, in loan origination, how can you manage leads more efficiently?
  • How do you integrate human decision making with the automation of mundane tasks?
  • You need to focus on the process first to make sure that it works right before you automate it. There is no benefit in automating a bad process. And, don’t just mimic the paper process…re-imagine the process, one step at a time.

Common Misconceptions of AI

  • The average person might not hear about AI in banking, but it can encompass a lot of things, and play an important role in the front and back office.
  • There is a perception that AI can be risky from a regulatory perspective. We need to build AI models that are transparent.
  • People think that AI will lead to job reduction. In reality, it will lead to job augmentation. It will free up people from the mundane tasks so that they can do more customer-facing, value-added work. In mortgages, we need high-touch with automation.
  • Some think that AI can run “unattended.” People need to understand that AI can only make decisions at a certain level. You need (human) support staff with the right skill set to work with and support AI.

Momentum in AI Deployments

  • Key advancements are taking place in document classification, machine vision and RPA. We need to show the ROI around these advancements and capabilities.
  • You need to identify the “quick wins” in AI deployments to show the rest of the organization the advancements that you are making. And, it’s good to “over communicate” these wins (and what you are doing, and why) to key stakeholders.
  • Rather than making a deployment too large, and therefore too difficult to complete successfully, automate one or two departments or processes first. After those are proven and successful, move on from there.
  • Previously, machine learning was seen as a “black box,” but now there are explainable AI solutions that are coming out that have auditable code. This provides greater transparency and is important for regulated industries.
  • Analytics will be a key area to watch in the near future – it will allow us to see what customers are deciding.

To hear more details and background related to the points above, listen to the MReport webinar.


Actionable Intelligence.Transform your business into a highly automated, competitive player.

Let AI Foundry show you how to increase the volume of mortgage applications you’re processing while simplifying manual processes and extracting the data you need—all while ensuring data accuracy.

Request a demo now

The executives on the panel included:

Manish Garg – VP of Product Management at Ellie Mae;

Nate Spiegel – VP of Process Improvement & Change Execution at Citizens Bank;

James Deitch – Co-Founder & CEO of Teraverde Management Advisors; and

Steve Butler – Founder and President, AI Foundry