Our Blogs

In 2020, AI & machine learning will boost Fintech by improving the quality and flexibility of payment, lending, and insurance services while also helping to find new borrower pools.

The traditional tech stacks of Fintech were not designed to anticipate and act on real-time market indicators and data quickly; they are optimized for speed and scale of transactions.

What is needed is a new tech stack that can be flexible and adapted in real-time to changing market and customer needs. AI & machine learning proves to be very efficient in interpreting and recommending real-world actions.

The following are ten predictions of how AI is going to improve FinTech by 2020, 

• Zest predicts that by understanding ML's operating expense (OPEX) savings, banks and other financial institutions can improve their business cases for AI pilots and production-level deployments. According to Zest, the development, validation, and deployment of credit risk models can reduce or reduce several recurring costs by switching to machine learning.
  

  By using modern ML tools to assess which data sources provide the most predictive power for a model, lenders can get the most out of their data acquisition spending. Also, lenders will switch to ML to simplify their IT and risk operations by consolidating into fewer models that can do the job of what used to be multiple linear models for each customer segment.

  
  
• Artificial Intelligence and Machine Learning are gaining critical mass in collections, providing insights into which approach to a given customer is most successful. For a few financial services firms, Zest has developed collection models and found them to be very successful. Collections logic is a good match with machine learning to predict which customers to wait on when bills are due past. 
  

  For example, with one bank, Zest found that ML models can accurately target borrowers with the highest probability of making a certain minimum payment based on the value of their loan within 60 days of their due date. Within three months, to forecast this repayment propensity of borrowers, Zest built two models from traditional credit offices and the bank's proprietary collections metrics.

  
  
• 2020 will be a break-out year for partnerships and co-option as payment, lending and insurance companies compete for a growing position in embedded financial services. The prediction of embedded fintech by Matt Harris of Bain Capital Ventures suggests a proliferation of cloud-based Fintech apps around the core: payments, loans and, insurance.
  

  This creates an ideal situation among incumbent lenders, start-ups, data aggregators and CRAs for AI-related alliances and partnerships. To Harris, the stack layers are based on communication, intellect, and omnipresence.

  
  

   

• To demonstrate the effectiveness of their strategies for talent management. Fintechs will increasingly adopt AI and ML in 2020 to find, recruit and hire the best candidates for development, engineering, marketing, sales, and senior management roles. In 2020, Fintech CEOs and CHROs will start upgrading programs for themselves and their teams to enhance AI fluency and mastery skills.
  

  Demonstrating the effectiveness of their talent management strategies. In 2020, Fintechs will increasingly adopt AI and ML to identify, recruit and hire the best development, engineering, marketing, sales, and senior management candidates. In 2020, Fintech CEOs and CHROs will start upgrading programs to improve AI fluency and mastery skills for themselves and their teams.

  
  

• Zest forecasts borrowers will increase the use of ML as a way of growing into the no-file / thin-file segments, especially through Gen Zers with little to no credit history. Traditional tech stacks make finding and growing new borrower pools difficult.

   
  
• Growth in the cost of compliance would decrease even more rapidly due to ML. Financial institutions that have AI / ML algorithms in the production record any change in a model and can generate all the necessary risk management documents in minutes instead of a compliance department that takes weeks to do it manually.
  

  Automated tools also shrink the time it takes to do fair lending testing by building less discriminatory models on the fly rather than the time-intensive approach of drop-one-variable-and-test.

  
  
• If a downturn occurs, ML will be blamed (although in a downturn it can actually help). Originally, this observation was made by Pankaj Kulshreshtha, CEO of Scienaptics at the Money 20/20 Conference held earlier this year. 

  Models built only in good times, when times go bad, can see their correlations break. Lenders who observe best practices when adopting AI and ML will ensure that their mod is stress-tested.

  
  

  

  
  
• Zest predicts that Fintechs will seek expertise in AI and ML modeling more than building their own expertise and teams, which will be more expensive and take longer. The future adoption rate of Embedded Fintech is based on how effective development efforts are today to minimize incidental bias and give more visibility to customers about how and why models deliver specific results.
  
  
• The adoption of AI by mortgage lenders to find qualified homeowners for the first time will increase as more realize Gen Z (23-36-year-olds) are the most motivated to buy a home. Long-standing hypotheses about first-time homebuyers and their motivations will change in 2020.
  
  
• In 2020, credit unions will implement ML to automate routine tasks and free human underwriters to focus on providing more customized services, including improving inquiry resolution and conflict management and fraud detection. Credit unions are based on an annuity-based business model that successively delivers higher profitability the longer a member is retained.

Credit unions will capitalize on ML by driving without added risk loan approvals and automating more of the loan approval process. By the end of 2020, 71 percent of credit unions plan to investigate, test, or fully implement AI / ML solutions, according to a Fannie Mae mortgage lender survey-up from just 40% in 2018. In order to improve investigation resolution, AI and ML will also be adopted across credit unions.