RESEARCH REPORT
RESEARCH REPORT
AI for 4 loan journey stages A typical loan application customer journey is comprised of four stages: processing, underwriting, closing and funding, and post-closing (see Figure 2). Prospective borrowers submit an application form, the lender reviews their finances and assesses risk, they receive a specific offer based on their risk, accept it, then close the loan and receive the money. Once the loan is closed, they receive follow-ups from the lender about payment terms and begin a new journey as a paying customer.
AI tools can be deployed across the four stages. Here are common AI solutions for each unique part of the loan journey.
Processing: This stage is very data-heavy with application forms and requests for detailed information. There is clear information with little need for human intervention. It’s the stage most primed for AI automation. Internally, AI tools can extract and validate data, verify loan completions, correct errors, and power real-time crediting decision-making to increase accuracy and speed at the beginning of the customer journey. Externally, virtual assistants can walk borrowers through steps and confirm accuracy and completeness of information to prevent delays. Underwriting: This stage is all about risk, which is again driven by data and clear rules. Here is where some decisions need to be made based on unique variables, so associate augmentation works better here in general. Customer-facing activities are minimal. AI tools serve employees here with AI-enabled risk modeling and scoring, decisioning support, and further verification, validation, and reporting. Closing: Here’s where the money comes in, so customers will want to know there are people behind the information for peace of mind, support, and trust. It’s imperative that companies provide layers of support and communication in this stage. Processes diverge from repetitive tasks to more personalized, unique interactions, making full automation unrealistic.
AI tools in this stage focus on optimizing the repetitive, impersonal tasks like compliance checks, document preparation, loan pricing and optimization, and data integration with title, escrow, and settlement providers. Customer interactions and explanations are best left to human associates to guide customers through the closing process. Post-closing: In the final stage, borrowers get their money, the deed is recorded, and title policies are issued. There is no room for error in this important stage, and the potential for extremely harmful fraud is high. People need to stay at the forefront of this stage, with minimal AI support through loan package completion, post-closing audit reports, and document archiving. Even in AI’s early days, there are AI solutions specific to the loan journey available. They include: Blend, DataRobot, FICO, Fiserv, Kasista, nCino, Socure, Upstart, and Zest AI (note: TTEC does not specifically endorse any of these vendors).
Figure 2: The 4 stages of bank loan application journey A typical loan application has four stages, each with opportunity to implement AI tools in support of the work within the stage.
+
Closing and Funding
Processing
Underwriting
Post-closing
• Review closed loan package • Review final copies • Underwriter decision review • Review closing conditions and instructions
• Evaluation of
• Regulatory
• Re-review loan package • Prepare closing package • Final conditions review • Pre-closing QC • Execution of
completeness of loan application
compliance review
• Verify credit
• Verify credit worthiness • Order third-party services like credit reports
worthiness and risk scores • Verify character, capacity, and collateral
“Feedback loops are a critical component of any augmentation. Being able to
closing documents and funds disbursement
incorporate the success or areas for improvement not only improves the AI running the processes, but also builds user confidence. This is the key to gaining adoption of technologies that augment – otherwise, associates and
AI Work
AI Work
AI Work
AI Work
• Risk modeling and scoring • Data verification • Reporting
• Document preparation • Compliance checks • Data integration
• Document archiving • Loan package completion • Audit reports
• Data validation • Integration with other systems • Virtual customer assistants
employees will ignore or find a way around what’s in place.”
– Aaron Schroeder, Director of AI Solutions, TTEC Digital
Source: TTEC
6 | AI/CX VALUE CHAIN ANALYSIS: BANK LOANS
7
TTEC.AI |
Made with FlippingBook Annual report maker