OptiFunder CEO Michael McFadden: What Mortgage Leaders Need to Know About AI
Michael McFadden is CEO of OptiFunder, St. Louis, Mo., a fully integrated and automated warehouse management platform for mortgage originators.
We’re inundated with AI claims and warnings, so what do mortgage leaders need to know about the benefits, risks and opportunities of AI? As the CFO of a large lender, predicting warehouse funding expense was challenging. The vast complexity of dynamic data needed for funding decisions made traditional methods impractical. I began exploring AI for a solution and ultimately founded OptiFunder. Now as a Mortgage Tech CEO, I have culled some great resources and hope to offer some helpful guidance for AI adoption based on personal experience.
With major, recent advances in Artificial Intelligence, awareness and use of AI is growing rapidly. AI’s ability to tirelessly factor myriad, dynamic datapoints from multiple sources instantaneously is being deployed for efficient problem solving.
As you evaluate a solution, consider:
• Why AI, what problem does it solve?
• What data sets is it considering? Understanding and controlling the data source(s) can improve solution confidence.
• How is AI used?
• Who created the solution—is it proprietary (built and utilized specifically for the task at hand), third-party and/or an open-source AI application? Does the company demonstrate expertise in this application?
At OptiFunder, we developed AI-powered algorithms to continuously compare multiple data points from our warehouse lender partners and lender clients to make strategic warehouse allocation decisions for superior and more predictable financial outcomes.
What is ML and how does it support AI?
Machine Learning (ML) factors changes in data over time to help AI make better decisions. For example, OptiFunder uses machine learning to continuously evaluate changes in measurable timelines, which informs the AI driving our optimization algorithms. When looking at AI, how are dynamic changes in data sets factored to continue to improve decisions?
What is optimization?
Optimization is a science that informs the best decision for a specific/strategic outcome. AI/ML unlocks the true opportunity of optimization. The savings we achieved were (and are) much greater than I would have anticipated, and more than people assume. When building OptiFunder, we created rules-based optimization algorithms that enable our clients to get the best (optimized) decision for warehouse allocations based on their strategic objectives (i.e., lowest cost of capital, highest ROE or specific funding targets). When the AI work is rules based, you have options to guide the desired results based on your needs—even as those needs change. How will the solution you’re considering change with your needs?
What AI risks should I consider*?
• Untraceable, Unauthorized AI Implementation: Your IT and Risk Management team should be consulted when considering enterprise software solutions; and your employees should be continually trained not to use or upgrade any software application without IT approval.
• Introducing and Magnifying Bias in Decisions: It’s important to question, know and trust the dataset informing your (human) decisions—and even more important when AI is making decisions. This article from MBA Newslink covers why (and important CFPB warnings) in depth.
• Personal Privacy Violations: As the datasphere grows exponentially, so does the risk of exposure—making personal data harder to protect. Data breaches/misuse can significantly damage a company’s reputation and present legal jeopardy. Understanding data sources, ownership, protection, and existing and emerging regulations is important.
• Lack of Transparency (“Black Box” algorithms): AI-powered algorithms are often complex and proprietary, which magnifies the importance of considering the other risk factors in this list and trusting the source of the AI-enabled solution you’re using.
• Unclear Legal Responsibility: What legal impacts could come from the AI you’re utilizing? With new technology comes new responsibilities, engaging the Legal and Risk Management teams in the review is advised, and using vendors demonstrating responsible practices and safeguards is paramount.
*For more details, read “What are the Risks of Artificial Intelligence.”
How do I know if I have a trustworthy solution-provider?
In addition to being sufficiently transparent, what should one look for in a SaaS or Tech partner? I recommend this checklist:
• Security, privacy and compliance (Vendor/Developer uses recognized standards and frameworks and System and Organization Controls.)
• Reliability and performance
• Support (training, maintenance, customer support)
• Vendor lock-in
• Integrations (APIs facilitate partner integrations)
• The vendor’s roadmap and finances (Are they consistently investing in product improvements?)
More info available from the list source here.
In addition to the list above, new solutions (AI-powered or not) should provide: Proven ROI, easy implementation (minimally intrusive, expeditious) and client endorsements. Thinking about building a proprietary solution, or customizing a current solution to meet your needs? This often requires staffing investments, inter-discipline focus and doesn’t take advantage of cost-efficiencies of a third-party solution. As you weigh your options, there’s never been a better time to consult not only your IT and Ops teams, but your Legal and Risk Management resources as well.
(Views expressed in this article do not necessarily reflect policy of the Mortgage Bankers Association, nor do they connote an MBA endorsement of a specific company, product or service. MBA NewsLink welcomes your submissions. Inquiries can be sent to NewsLink Editor Michael Tucker at email@example.com or Editorial Manager Anneliese Mahoney at firstname.lastname@example.org.)