· Connectors · 3 min read
Large Language Models in Finance: Executive Summary for Board Members
Large language models are transforming key financial processes:
- Automated identification of embedded leases improves ASC 842 compliance
- Enhanced stress testing and risk management capabilities
- Streamlined revenue recognition under ASC 606 and IFRS 15
Board implications include improved financial reporting accuracy, enhanced risk management, and potential for significant operational efficiencies
Generative AI can be used to create stories about cats on spaceships, summarize research papers, and come up with terrible jokes. But large language models are also capable of reshaping core financial processes; accordingly, it’s a relevant area for board members who oversee financial governance and risk management. I recently wrote a post for 273 Ventures (my legal AI company) where I explored a few financial use cases for LLMs: Unlocking the Potential of Large Language Models in Financial Use Cases. I’ve included a high-level summary of the use cases below, as well as some related board considerations:
Embedded Lease Identification
- AI can automatically identify embedded leases in various documents (service agreements, data center contracts, etc.)
- Improves compliance with ASC 842 standards
- Reduces risk of overlooking hidden lease obligations
Board Consideration: What kind of structural inputs are required to shift to this type of process? Does our organization have sufficient capital in the short-term to support this type of transformation?
Stress Testing and Resolution Planning
- AI enhances analysis of vast amounts of contractual data for stress testing scenarios
- Improves identification of subtle contract nuances and potential risks
- Supports more comprehensive and efficient regulatory compliance (e.g., Dodd-Frank Act requirements)
Board Consideration: The use of LLMs allows for better capture of unstructured data for use as model inputs. Can we leverage this technology to strengthen our risk management framework and regulatory compliance in other areas of the organization as well?
Revenue Recognition
- AI streamlines the process of identifying performance obligations and transaction prices across scattered documents
- Supports compliance with ASC 606 and IFRS 15 standards
- Improves consistency and accuracy in revenue recognition practices
Board Consideration: What are the implications for our financial reporting processes and potential for reducing errors or inconsistencies? Is strategic or operational change management required for us to capture the value of using AI for this? If you’re on the audit committee and want to take a deeper look at revenue recognition using AI, I have another post on that topic.
Strategic Implications for the Board
Enhanced Accuracy and Efficiency: LLMs have the potential to significantly improve the accuracy of financial reporting while reducing the time and resources required. Consider how this might impact our financial operations and resource allocation.
Risk Management: The improved capabilities in stress testing and contract analysis could lead to more robust risk management practices. How should we integrate these AI-driven insights into our overall risk governance?
Compliance and Auditing: As AI takes on a larger role in these processes, we need to ensure that these systems align with regulatory requirements and can provide transparent, auditable results. How do we maintain appropriate oversight?
Talent and Skill Sets: The adoption of these technologies may require new skill sets within our finance and risk management teams. What strategies should we consider for upskilling existing staff or acquiring new talent?
Competitive Advantage: Early adoption of these technologies could potentially provide a competitive edge in operational efficiency and risk management. How does this align with our overall strategic goals? Are there other areas of the organization that would recognize greater efficiency gains that we should focus our resources on first?