CONTESTED CASES: SERVICERS, DO YOU KNOW WHO YOUR “AI-PMKs” ARE? HERE’S WHY YOU SHOULD.

---

Fresh from USFN Industry Forum here in Denver, it’s no secret many believe (myself included) we’re headed for an impending default wave.  The combination of rocky economic indicators, affordability pressure, sleeping risk in escrow accounts, increasingly frustrated (and emboldened) borrowers/consumers means we’re at risk for a higher rate of contested cases in default servicing.  Add widespread AI implementation in consumer finance, and servicers will certainly see challenges to its use in anything default and foreclosure-related.

If you believe “the best defense is a good offense,” servicers and their counsel have a vested interest in developing either a single expert or a series of experts who can clearly identify and explain: (1) where AI is being used in default servicing operations, (2) how it is being used and (3) the soundness of the outputs. 

Servicers need to develop their roster of AI experts now. I’ll share how I think about this as a defender of servicer interests.  Where a company or other organization is a defendant, Rule 30(b)(6) of the Rules of Civil Procedure allows a plaintiff to identify a list of topics for inquiry at deposition rather than naming a specific individual to testify.  This rule is based on efficiency. The idea is, the plaintiff has no idea who the experts are, so it’s the responsibility of the corporation to designate one or more persons capable of testifying knowledgably on the company’s behalf regarding the subjects identified. These depositions are fairly common and are often referred to as “person most knowledgeable” or “PMK” depositions. This is a helpful litigation paradigm to use tactically in bracing for what’s next in contested default cases. With servicing-based AI, the roster of PMKs might include the CIO, CISO, and IT director level employees within your organization, as well as their counterparts and the CTO or default product executives within vendors of a servicer.

With that in mind, as a mortgage servicer, do you use AI in customer service? In an environment where meaningful assistance to borrowers in distress will be paramount, do your bots and assistants really provide the needed assistance? How do you know they work? How do you ensure they don’t put borrowers on a hamster wheel of useless information?

Do you use AI to monitor your portfolios for payment patterns or to develop queues for outreach? Given the rising sensitivity to data mining, how do you know your solicitation efforts are soundly based?

Do you use AI in any aspect of loss mitigation such as information gathering, income verification, or decisioning? With many emerging state laws focused on AI use in any aspect of decisions that materially impact a consumer, how do you know those functions operate as intended and the relief options offered are appropriate?

If you are a servicer and I am defending you, I will need unfettered access to professionals in your organization, or within your vendor organizations, who can clearly articulate, at a minimum:

• Where and how is AI used in the default servicing process

• How AI is governed and how the AI at issue was vetted and approved internally prior to putting it into production

• What data are used in the AI model

• How the model was designed, trained, refined and tested (and the evidence documenting that work)

• How the model has evolved since go-live (and the logs evidencing changes)

• How employees are trained on AI in their area

• How is the operation of the model and its outputs monitored and how are the monitoring results reported

• How are issues with AI detected, documented, remediated and cleared

• Where are the humans in the loop and how knowledgeable and trained are they

• Would the outcome being challenged have been the same if it had been done manually by a human being

  
  

Looking at this list, how ready are you to describe and defend your AI-enhanced default servicing processes? Do you know who your “AI PMKs” are? Are they within your ranks or with one of your vendors? Are they spread across different companies or across departments? Do they understand their piece of the equation as well as how their domain fits into the overall picture of governance and use of AI in your default operations as a whole? Does your leadership understand these issues and recognize the AI PMKs?

While this may all feel like a lot, keep in mind that

you can meet impending AI-based default servicing challenges defensively or offensively.  I prefer the latter. Knowing and naming your AI PMKs puts power in the hands of the servicer because you designate these people to represent your AI program if and when the time comes.

If you haven’t given thought to who your AI PMKs are, there is no time like the present. The Mortgage Law Firm is committed to the success of its mortgage servicing clients in this rapidly evolving space. Those cases will be coming and as the lawyers closest to the issues The Mortgage Law Firm is prepared to step in with seamless in contested cases challenging AI use in the default space.  Let’s make sure your playbook is ready. That is what will put us both in the best position to represent your interests successfully.