Blockchain, Artificial Intelligence & Machine Learning: Hype or Help?

All Three Could Contribute To Faster Digital Transformation

An illustration of connected blocks in a chain, blockchain

Insider

The mortgage industry has been talking about ending its ‘Paper-Palooza’ for at least 20 years as we linger behind other industries like healthcare and insurance. While presently enjoying a surge in innovation and investment, it’s just a small spark. Digital technology opportunities span the entire ecosystem, bound only by the willingness of its participants.

Artificial intelligence (AI) and machine learning (ML) are the most understood and deployed, thereby leading the way in these early stages. Blockchain, on the other hand, is more “fuzzy” to many, yet has a persuasive cast of evangelists.

There are companies, and even countries, being built on blockchain tech, such as Figure and Liquid Mortgage. Estonia is blockchain crazed, starting with its own government entities, promoting exponential business growth. They apparently began 20 years ago. Sweden and the UAE are said to be leaders in mortgage blockchain.

In U.S. mortgage, there appear to be more theoretical blockchain use cases than actual deployments, for now. Although, when you understand what is possible and some current successes, optimism abounds.

Artificial Intelligence (AI) & Machine Learning (ML)

Companies across our industry are increasingly deploying AI/ML, with big gains surrounding borrower engagement, credit decisioning, risk and portfolio analysis and fraud detection – to name a few.
A great barometer of adoption is the corresponding level of regulator interest. Recently, the Federal Housing Financial Agency’s (FHFA) Division of Enterprise Regulation (DER) has increased monitoring of AI/ML usage at Fannie Mae and Freddie Mac. AI/ML can pose added risks including operations, compliance, modeling and financial.

In February 2022, the DER issued an advisory bulletin specific to the use of AI/ML, said to be the first such guidance from any mortgage regulator. The guidance asks institutions to increase risk management practices. The FHFA, as well as the Mortgage Industry Standards Maintenance Organization (MISMO), also recommend institutions define a code of ethical standards for checks and balances. With adoption already well-underway at the government sponsored enterprises (GSEs), AI/ML has arrived.

Advances in AI/ML are happening so fast some leaders concede they struggle to keep up, causing a bit of decision paralysis. According to the multiple senior leaders we spoke to, mostly representing Top-40 originators, some said they feel overwhelmed by the pace and breadth of digital tech. What a strange dichotomy – we waited for years to begin, and we now slow-walk action on where to start. It is a fantastic disposition, unless the result remains indecision.

Given the steady growth of experienced AI/ML vendors and deployments, this technology will soon become compulsory, and those institutions that delay will be disadvantaged on many fronts.

What About Blockchain?

Too few of us thoroughly understand blockchain. However, experts speak with great confidence about how it will someday obliterate the mortgage world as we know it. We certainly believe the vast potential, in that nothing is poised to revolutionize more than blockchain. There are just some immense complications, far more complex than with AI/ML.

You can spend days researching articles, white papers and eBooks about blockchain in mortgage. It is equally promising and captivating when you envision the potential for:

Originating in days, not weeks or months
Registering and storing property/title/deeds on the blockchain, vs. title insurance
Single step auto-verify/QC, versus multiple duplicate efforts/parties
Near real-time payment, settlement and reconciliation
Auto-alerts and reports on credit, compliance, risk, etc.
Faster MBS & MSR transfers and liquidity, less instability
Eliminates countless middlemen and redundancy
Radically drives overall costs down (efficient, accurate, compliant, etc.)

Our excitement, however, tends to normalize when we ponder mindset changes required. Blockchain technology is so robust, with such radical change potential, one must wonder how executives will embrace ‘everything’ (i.e., info, records, data, payment, etc.) on a blockchain platform described as:

A public distributed ledger for storage of info/data/actions in blocks on a decentralized database with complete transparency. No one person/company has sole control. Every transaction is sent to every other ‘privileged’ party with approved access.
When a block is added to the network, multiple pre-defined parties must validate and approve it before it gets added to the ledger.
Smart contracts, in simplistic terms, enable rules and task execution for certain events/blocks. Example: Notify all three interested parties when a task completes.
Lenders must have trust in block (data) creators and accept they may not own it as they always have in the past.
Documents get validated and assigned a unique code for the block, called a hash. Only the hash and related smart details (i.e., verified, QC’d, compliant, etc.) post to the ledger, with physical docs going to a separate eVault. This doc never needs reviewed again later in the lifecycle, instead trusting the original validations.
The mortgage process includes many participants. For a lender to achieve any blockchain scale, most other engaged parties must also someday deploy

At this point, there are probably at least a few readers thinking more about ‘block-CHANGE.’ Understandable – we are a highly nuanced and risk averse industry. Flipping that switch, no matter how many years in the transition, is intimidating.

All the references to “public, decentralized and transparency” beg the question of security. Blockchain pundits generally claim it’s at least as secure as today’s platforms, if not more. Each block references the previous block, forming the chain. If a fraudster tries to make an unauthorized change, the links with other blocks are broken.

This, in turn, invalidates the chain from the change point. Because it is a distributed ledger, the fraudster would have to replicate the exact same change within every single node instance. While not impossible, fraudsters will likely be more apt to pursue easier targets.
Blockchain challenges are undeniable and run far deeper than described herein. Fortunately, the same can be said about the huge upside. Success requires a modest starting point with many deployment phases, being a multi-year marathon, not a sprint. Change management protocols will be equally as essential as project management.

Faster Digital Transformation

The real question is – what will drive faster digital transformation, be it AI/ML or blockchain?

Triggerable events are usually business plans with hefty top-line growth objectives or bottom-line budget challenges. The larger the perceived gaps, the greater automation interest. Competition and compliance issues will also motivate leaders. Regardless, guarding the customer experience will always remain an imperative.

So, is widespread digital transformation hype, or help? We have settled on hope. Small sparks possess wildfire potential.