Institutional Loan Origination

Institutional loan origination is a complex, multi-party process that involves extensive manual effort in document drafting, tracking changes, and data entry. Despite some automation, inefficiencies persist due to highly customised loan agreements, lack of integration with workflow tools, and reliance on manual validation. This whitepaper explores these challenges and investigates potential solutions to enhance accuracy, efficiency, and scalability.

Through global research and industry analysis, DB Results has identified solutions both within and outside the banking sector. Emerging technologies such as low-code platforms, AI-driven document analysis, and Robotic Process Automation (RPA) offer viable solutions to streamline a bank’s institutional loan origination process. By leveraging automation for document creation, change tracking, and data extraction, banks can reduce costs, minimise errors, and accelerate execution even with highly customised documents. This paper presents potential solutions to modernise processes, while providing flexibility in loan structuring, and maintaining compliance obligations.

Current State and Challenges in Institutional Loan Origination

The institutional loan origination process is inherently complex (Figure 1), characterised by customised and extensive loan documents, which can often run into hundreds of pages. These documents pass through multiple parties, including customers, sales teams, credit risk teams, external lawyers, and operational teams, before final execution. The multi-party involvement and non-standard loan document structure introduce a range of challenges (Figure 2):

• Loan Document Drafting: While there is some automation in the construction of loan documents through selectable standard terms, the overall drafting process remains largely manual because of the level of customisation in loan agreements. This approach is resource-intensive and prone to errors, particularly for large and complex agreements.
• Tracking Changes: The negotiation and finalisation of loan documents involves multiple rounds of revisions, which are typically not systematically tracked. This leads to inefficiencies and potential errors when finalising the loan terms.
• Final Document Comparison: Before execution, the original issued loan document must be manually compared against the final “accepted” version to capture changes made during the execution process. This step is expensive, time-consuming and prone to error.
• Post-Execution Data Management: Once loan documents are executed, critical details – such as facilities, facility limits, conditions precedent, and securities – must be extracted and entered into other systems. This step is predominantly manual with limited automation, reducing operational efficiency.

Typically, the automation tools in place, such as those for document construction and data extraction, are custom-built in-house. These bespoke systems often:

• Have Narrow Functionality: They address only specific aspects of the loan origination process.
• Lack Integration: There is minimal integration with broader automation or workflow tools, preventing end-to-end efficiency gains.

Our Approach

Our research spanned three streams that focussed on addressing these 5 key pain points:

• Global desktop research to assess how multiple banks’ loan origination processes operated, including platforms and tools.
• Engaging our partners (UiPath & OutSystems) to identify case studies where they have implemented automation solutions in banks.
• We also looked outside the banking sector (specifically the legal profession) where processes, tools and platforms used in these sectors could address some of the pain points identified in the institutional loan origination process.

Industry Solutions

Automation and straight-through processing appear to be well-established in Retail / personal loan origination operations; however, it is far less common in Institutional operations due to the length, complexity and variability of Institutional loan facilities and their corresponding documentation.

Most scenarios where higher levels of automation were observed occurred where banks had completed an end-to-end transformation of their loan origination, administration processes and core banking platforms. Many banks are implementing low-code solutions coupled with Robotic Process Automation (RPA) and/or AI. This approach enables organisations to realise value (and iterate) significantly faster and at lower cost than traditional implementation methods, including Commercial-Off-The-Shelf (COTS) solutions.

Pain Point 1: Manual Creation of Loan Documents

Few solutions fully address the ‘manual’ construction of loan documents, given the amount of variability allowed in such documents. However, further review of loan documents to fully understand the amount and location of variability would be beneficial. AI / RPA can be used to accelerate the investigation (Figure 3), to identify additional clauses that could potentially be “standardised” (with no variability and standard variations) to reduce the amount of manual effort in the construction of loan documents.

Pain Points 2, 3 & 4: Tracking and Approving Changes to Loan Document Versions

The negotiation, tracking and approval of changes to loan documents is not unique to banking – it is a contract management process. Solutions used in the legal profession to manage the approval and execution of legal contracts (specifically Contract Management Systems) provide many features that could streamline a bank’s process to negotiate, update, finalise and execute loan contracts, including:

1. Automated Document Assembly: AI-driven document creation reduces time and enhances compliance. NOTE: The investigation outlined above is still required to identify clauses that can be standardised, however, could be included as part of the implementation of a Contract Management System.
2. AI Contract Analysis: AI tools organise and extract key contract data for improved outcomes.
3. Process Automation & Workflow: Streamlined approval and signature workflows to accelerate execution.
4. Document Management & Search: Secure storage with powerful search features and version tracking.
5. Electronic Signature Integration: E-signatures simplify approval and reduce paperwork.
6. Intuitive Collaboration: Collaboration tools ensure everyone works on the latest document version.
7. Contract Portfolio Reporting: AI and data visualisation improve decision-making and risk management.
8. Integrations: Integrates with CRM and other systems to streamline workflows and provide unified data views.
9. Data Security: Robust security protocols, encryption, and cloud hosting with 24/7 monitoring.
10. Mobile Capability: Mobile access allows contract management on the go, enhancing flexibility.

Importantly, licensing is relatively inexpensive (a few thousand dollars per month) and many are offered on a secure, software-as-a-service basis.

DB Results has built a similar solution for a customer. The system controls the contract negotiation and execution process with secure access for all stakeholders, a virtual data room, real-time updates (by all parties), and document version control. This solution has significantly reduced the time and effort to track changes during the contract negotiation and execution process, and the version control and real-time collaboration capabilities make handwritten updates unnecessary.

Other applications specifically designed to compare legal documents also exist, providing lower cost alternatives to end-to-end Contract Management Systems. These solutions typically improve the latter stages of the process (pain points 3 & 4).

Pain Point 5: Loan Data Extraction and Entry to Bank Systems

Automated data extraction and entry solutions are used extensively in banking and many other sectors. Whilst it is more challenging with Institutional loan documents, given their complexity and variability, best-of-breed RPA solutions have developed sophisticated, AI-supported, document understanding capabilities that can read and extract data from documents of varying length and content. UiPath’s latest version of its Document Understanding framework contains pre-trained machine learning (ML) models that can process less structured documents with high levels of accuracy and minimal setup time.

The scope of future state automation (versus current solutions) needs to be understood before committing to an enterprise-grade RPA solution. Notwithstanding, the benefits of implementing an RPA platform across multiple use cases in a banking environment can be significant (Appendix 3).

Globally, the combination of low-code/no-code platforms, AI, and RPA is driving efficient loan origination and contract management. By looking outside the banking sector and taking advantage of the latest advances in AI-driven document analysis and RPA, banks could implement cost-effective solutions that significantly increase productivity in their Institutional Operations team and reduce the time to document, execute and set up Institutional loan facilities for their customers.