What Are The Challenges To AI Adoption In Lending? 8 Experts Share Their Insights

  • 30 September 2019
  • Sam Mire

Prevailing trends tell us that AI is becoming a real asset in the lending industry, but it's not quite a necessity. So what is stopping investors from seeing artificial intelligence as a tool that they can't do without?

These investing industry insiders shared their views on the greatest hurdles facing AI adoption in their field. Here's what they said:

1. Dr. Marlene Wolfgruber, Director of Product Marketing at ABBYY

“The goal of applying machine learning and AI is to provide the best possible outcomes and the most efficient process. Many lenders think they know how their lending process performs but it is often far from reality. The number one challenge for lenders here is keeping track of the performance throughout the whole process chain – from origination to underwriting to funding – in order to be able to improve the process and increase customer satisfaction.

It is a critical pain point as organizations start to realize there are blind spots and they may need to diverge from their original processes. However, new applied machine learning and AI technology, called Process Intelligence (PI), enables lenders to analyze less structured processes, identify opportunities for improvements, increase the speed and accuracy of executing the lending process and reduce costs.”

2. Keren Moynihan, co-founder of Boss Insights

“The financial institution has to be willing to provide data and the tech company has to be able to show a secure source.

There are 3 big challenges:

1. A lack of access to a holistic, timely, and accurate source of data.

2. Privacy concerns – In terms of how data will be used and what the impact will be on those sharing it.

3. Awareness – Many firms employ AI to assist in business processes,
but it remains a trade secret among a select few – and that group is getting a competitive edge.”

3. R.J. Talyor, founder and CEO of Pattern89

“In an eMarketer survey, 32.9% of marketers have reported that applying AI to their current roles and workflows is their biggest challenge. 30.6% say they are unclear about what AI is and how it works, and 28.5% say their budgets can’t afford it.

For those marketing AI products, it is necessary to explain that AI won’t threaten jobs, but it will change them. Incorporating AI into advertising will allow advertisers to focus on creative and strategic work, while their software completes smaller tasks and analyzes data. Educating people on what AI is and how it really works is the best way to combat these challenges.”

4. Jeff Silberman, Counsel in Reed Smith’s Financial Industry Group

“Implementation is the greatest challenge to AI adoption in lending. In the abstract, the benefits of AI are obvious to most CIOs, and many lenders are aware that the technology is quickly becoming necessary. But there can be a natural push/pull that occurs when rapid technology use cases are clear for product or data owners, but not necessarily clear from a documentation and liability management standpoint for internal counsel.

It seems like one of the keys to success for adoption is for the user to identify a specific area of application and not try to tackle too much too quick. The initial AI-based project should serve as a stepping stone to broader initiatives that will have greater impacts on the organizations.  Another key to success is, of course, working with transactional counsel that understands the technology, the use cases and the appropriate contractual terms for the solution being implemented.”

5. Douglas Merrill, CEO and founder of ZestFinance

“The number one challenge is typically the complexity of implementing AI in existing IT. Those complexities are rapidly dissolving as more banks adopt open-source ML modeling tools and infrastructure that fits with the technologies they already run. The other big challenge for adoption is the black-box dilemma of most ML models. That’s what we’ve spent the last two years solving. Lenders have to know whether their model is unfairly impacting protected classes (race, gender, age) or making decisions that will hurt their business. It’s essential to be able to interpret both the inputs and outputs of an AI/ML model to ensure it’s not perpetuating bias or going off the rails.”

6. Steve Comer, Director of Financial Services at Hyland

“The seemingly obvious answer to this question would point to the cost of implementing AI solutions and the time to ROI. But a bigger challenge points away from the technology itself and questions whether or not regulations will be able to keep up with the advancements in technology. As big data continues to become bigger data, AI tools will have the ability to draw from data points that weren't fathomable even five years ago. The prevalence of data today would conceivably allow for higher quality lending decisions, but regulations will have to be continuously rewritten to very clearly define what access to data is considered acceptable to the strengthen the quality of the loan process vs. an invasion of privacy.”

7. Anis Uzzaman, CEO of Pegasus Tech Ventures

“We believe the #1 challenge is that a large part of investing is making decisions about people and revenue. As of now, we are not aware of widely available AI solutions that can provide contextual feedback about these factors.

AI can potentially be helpful in providing a digital snapshot of people based on the information available online, but we have not seen a solution to provide contextual information related to a team’s ability to execute a given idea within a particular target market. In addition, while Microsoft is rolling out new pattern recognition technologies for excel, there is still no known solution to incorporate micro and macro market conditions relative to revenue traction and projections.”

8. Joshua Jones, CEO of StrategyWise

“Regulations and ethics. AI and advanced analytics are great for getting you to a high confidence level, but many times being 90% confident that you’re not doing something, like unfair discrimination, is not enough. Finding that right balance, and ensuring that appropriate check points, validations, and reviews are in place is a key success factor.”

Have expert insights to add to this article?

Share your feedback and we'll consider adding it to the piece!


About Sam Mire

Sam is a Market Research Analyst at Disruptor Daily. He's a trained journalist with experience in the field of disruptive technology. He’s versed in the impact that blockchain technology is having on industries of today, from healthcare to cannabis. He’s written extensively on the individuals and companies shaping the future of tech, working directly with many of them to advance their vision. Sam is known for writing work that brings value to industry professionals and the generally curious – as well as an occasional smile to the face.