Back to Industry Trends & Insights

AI in Lending: How to Implement AI so it Works for You

Social Journey

In order to succeed in using artificial intelligence, it pays to start small, stay focused and make sure that you take your team on the journey with you.

When you first introduce AI into your system, you do it because you’ve noticed a problem. The key in the beginning is to start small and stay focused and solve that problem.

Because it’s easy to suddenly see all the potential at your fingertips, many people make the mistake of trying to solve everything at once, and end up solving nothing. It’s far more effective to identify one area, get complete buy-in from the staff who are directly impacted by the new processes and learn some lessons before moving onto the next phase.

The first broadcast of the new [email protected] web series called A Practical Approach to AI in Lending featured a dialogue between AI Foundry’s Alex Levi and Kyle Manseau, the SVP Operations of the Allied Mortgage Group, and the man responsible for spearheading this very exciting project.

Kyle spoke at length about the process of implementing AI Foundry’s Agile Mortgage software; about how successful it has been, how well the transition was managed, and how his staff have successfully adapted to using the new tools.

For Allied Mortgage, the results have been impressive. Since implementation, they have run more than three thousand loans off the system. “Without artificial intelligence, it would have taken between six hundred and fifty and seven hundred man hours we're now doing in twenty five man hours. That's huge for an organization of our size.” Manseau said. 

Allied Mortgage Group took the decision to first deploy the Document Indexing & Filing robot to accelerate processing and reduce costs, and remove the possibility of human error.

In practical terms, Allied wanted the process to be easy enough “so that the UPS guy could deliver our closing package, put it on the scanner and walk away, and then it would be done.

And we've accomplished almost entirely that.” 

Arvind Jagannath, AI Foundry’s Director of Product Management agrees strongly with that kind of a focused approach. “Rather than trying to overhaul an entire business process by implementing AI on an end-to-end basis,” he explained, “ a better approach may be to pick one specific function to automate, and then apply AI and automation to that function only. Then, after your organization gets comfortable with how to implement it and realizes the benefits and ROI, you can move on to another specific function.”

“Scope drift is the kiss of death for many tech programs.” says Kyle Manseau, the SVP of Allied Mortgage Group, when he refers to projects that require continuous expansion and a changing mission.  “You can’t take your eye off the ball. You have to identify a single error in your ops and stick to it.”

Allied Mortgage decided to start with its post-closing operations.

Post-closing is performed after mortgage documents have been signed and returned to fund a loan. The process involves rigorous auditing and reporting to ensure all documentation is accurate and in compliance with underwriting and loan processing rules.

The cognitive robots assist lenders with the management of loan files before delivery to respective custodial service providers.

Utilizing these cognitive robots helps the lender to:

  • Ensure receipt of signed and funded loan documents from various loan stages
  • Review legal, origination and HMDA documents for regulatory compliance
  • Conduct loan data integrity checks and provide quality control reports stating any exception

These types of process-driven and meticulous quality control tasks are vitally important for a business but are usually very under stimulating for the staff who have to perform them.  It’s simply not the best way to make use of your human resources, while, thankfully, it’s the perfect workspace for today’s bots.

The bottom line is, when it comes to AI and software robots, you do not need to “boil the ocean” and overhaul your entire back office all at once,” AI Foundry’s Alex Levi emphasizes. “A phased approach with small steps works best, giving your staff time to learn how to implement and support AI in your systems.”

Manseau told AI Foundry that ‘It wasn't us taking something from you and then saying “how do we put it into a process within our operations. No, we looked at where we were and said now here's where we start. Your approach gave clarity on identifying the inefficient area that we wanted to change and what the software could do for it.’

How To Better Manage Your Staff

While you implement artificial intelligence in your system, you need to be thinking carefully about how to deploy all the extra time that your employees will have. They should not feel that they have been replaced, but rather that they have been let loose to do something more creative or more useful.

Allied Mortgage understands that the industry is cyclical, and they’re trying to break out of that. “We know we’re going to be busy in Q2 and Q3, then slower towards the end of Q4 and back into early Q1.Traditionally, we had to ramp up and bring in staff and train for the busy times and then let them go in slow times. We didn't want that reputation of hiring and firing.”

Instead the company decided to automate the menial tasks, deploy some resources and bring their people into other areas.

“We're able to take the talent and disperse them, and we didn't have to double staff in the post-closing task. With AI, we can double tomorrow and not have to add heads. So that’s a game changer in efficiency and cost-effectiveness for our operations. “ Manseau told Alex Levi. 

What Do The Bots Do?

One of the most useful aspects of working closely with a customer is that you get to see in real, practical ways how AI technology impacts a real business. “The mortgage industry clearly needs automation as loan processing costs are rising every year, loan defects remain stubbornly high.  We are really excited to see our customers deploying them in production,“ added Steve Butler, President of AI Foundry.

Manseau walked Levi through a typical interaction. “In post-closing, the bot takes a package from Fedex envelope and they walk away. We scan it, we log in at the end of the night, the bot ID’s the loan and all the documents within it, and does a quick cursory review to make sure it's a full package. It then completes the task and that automatically goes into Encompass, It indexes everything into the correct file folders, populates our post-close and makes it ship ready within compass. Then the loan is ready. And it can do that more than 35 times a day.”

Before, when Allied Mortgage got 30-35 applications, they couldn't get through all of them in one day. Now the bots have it all ready for capital markets scheme to trade the same afternoon.”

Manseau concluded by saying “We have become enamored with this tech and I truly believe it will change the world.”

Want to hear more?

Sign up here to access the Practical approach to AI in Lending webinar.

Get in touch today to find out what our bots can do for your lending.