LLMs and GenAI in Audit
In this blog post, we focus on potential applications of Large Language Models (LLMs) and generative AI (GenAI) in Audit. The objective is to introduce this transformative technology and discuss how it can empower professionals whilst raising awareness of the associated risks of implementing GenAI in day-to-day practices.
Large Language models and Generative Artificial Intelligence
LLMs are advanced AI models trained on vast amounts of text data which utilise neural network architecture and natural language processing models. It is designed to process and generate human-like text. What we commonly refer to as GenAI, such as ChatGPT, Gemini or Claude, are trained with vast amounts of textual data collected from various sources to perform a wide range of tasks from answering questions to generating coherent articles. Other examples include text-to-image models like DALL-E and Midjourney, which help users create image content. As such, GenAI helps its users create content including text, images and audio.
In essence, LLMs are a specific type of GenAI which focuses on language tasks while GenAi is a broader category that includes various forms of content generation across different modalities.
LLMs in Digital Transformation
As GenAI continues to evolve, its potential to revolutionize various industries becomes increasingly evident. The examples below provided by CPA's report highlight how leveraging LLMs and GenAI can streamline operations, enhance decision-making, and drive innovation. By automating complex workflows, improving customer engagement, and optimizing product development, these technologies not only address current challenges but also unlock new avenues for growth. The future of digital transformation is intertwined with GenAI, positioning it as a critical enabler for businesses to stay competitive and responsive in an ever-changing landscape.
- Automating workflows from beginning to end
- Handing off numerous tasks to digital assistants that function as resources within the organisation or for clients
- Hyper-personalising marketing content, taking on routine sales tasks and conducting better lead identification to optimize business development and growth.
- Allowing companies to design, test and bring new products to market more quickly.
- Speeding use of tools, such as 3D printing and computer-aided design, and identifying and enhancing design options.
- Minimising accounting profession’s staffing shortage by taking on a wider range of tasks, even beyond the repetitive steps that many AI tools are already handling.
- Analysing and identifying potential instances of fraud and other risk anomalies.
- Potentially making it easier to address environmental, social and governance (ESG) issues.
Use cases in Big 4 audit firms
The Big 4 audit firms are leading the charge in integrating GenAI into their operations, each with a unique approach that highlights the transformative potential of this technology. From Deloitte's cognitive tools that streamline document review to PwC's significant investment yielding productivity gains, and KPMG's ethical AI framework, to EY's enhancement of audit quality through AI, these firms are setting new standards in the industry. As GenAI continues to evolve, its role in reshaping audit processes will likely expand, driving greater efficiency, accuracy, and innovation across the sector.
- Deloitte has internally developed GenAI based cognitive technology to assist their employees with their existing workflows. This technology automates document review which allows practitioners to evaluate an entire population of contracts to identify key information, missing information and tracks subsequent revisions overtime.
- PwC has invested $1 billion investment and their in-house analysis claims to see 20% to 40% productivity gain. PwC is using GenAI in their IT development (i.e., software development process), finance data analysis, document summarisation and generation, and creating marketing contents.
- KPMG have launched “Trusted AI” framework to leverage the potential benefits of AI. The Trusted AI framework is designed to help KPMG’s client with GenAI solutions adoption in a responsible and ethical manner to mitigate potential unintended consequences of AI adoption in different business processes.
- EY adopted AI technology directly into its audit process. Specifically, they use machine learning to process, analyse and extract data from contacts, invoices, and images to improve the audit quality as well as providing additional value for its clients.
Prompting GenAI to work for you
Like any teamworking experience, effective use of GenAI requires a clear communication. GenAI’s usefulness hugely depends on clear prompt and guidance from the users. This exercise often refers to as ‘prompt engineering’. Six Principles for the Effective Use of Artificial Intelligence Large Language Models - The CPA Journal, provides numerous examples of how prompt engineering can improve users’ experience with GenAI as well as six principles of effective application of GenAI. They are as follows:
1) Develop specific questions and avoid broad requests
2) Actively monitor the inquiry and interpretation process by breaking down complex tasks into verifiable subtasks
3) Understand LLMs’ contextual boundaries, and do not input private, sensitive, or proprietary information
4) Carefully scrutinise and recalculate quantitative responses
5) Rely on other sources for factual information, especially for less prominent topics
6) Use LLMs to enhance rather than replace human expertise.
The future of Audit with LLMs
The IAASB published its strategy and work plan for 2024-2027 to enhance consistency and quality of audit and assurance standards worldwide. The work plan reflects the crucial role of audit and assurance in fostering trust in the world’s economies.
Key highlights of the Strategy include:
1) Strategic Focus: The IAASB aims to enhance the consistency and quality of audit and assurance standards worldwide. Their goal is to strengthen trust in global markets and improve the quality of engagements.
2) Priority Projects: The IAASB plans to complete several priority audit and assurance projects, with particular emphasis on fraud, going concern and sustainability assurance.
3) New Initiatives: Some key new initiatives and projects include:
· Supporting adoption and implementation of the overarching standard for sustainability assurance engagements
· Establishing an IAASB Technology Position
· Conducting post-implementation reviews
· Standard setting on audit evidence, risk response, materiality, and reviews of interim financial information
4) Collaboration: The IAASB emphasises collaboration with stakeholders across the external reporting ecosystem, including the International Ethics Standards Board for Accountants (IESBA), regulators and other standard setters.
5) Addressing Technological Changes: The work plan includes a broad-spectrum update of the ISAs to address the impact of technology, particularly focusing on the ISA 500 series related to audit evidence.
6) Standard Revisions: Several specific standards are slated for revision, including:
· ISA 320 on materiality
· ISA 330 on auditor's responses to assessed risks
· ISA 620 on using the work of an auditor's expert
· ISA 720 on auditor's responsibilities relating to other information
· ISRE 2410 on review engagements
7) Balancing Specificity and Principles: The IAASB is working to strike a balance between providing specific guidance and maintaining principles-based standards to avoid a checklist approach that could potentially reduce audit quality.
This strategy reflects the IAASB's commitment to evolving their standards to meet current challenges in auditing and assurance, particularly in areas like sustainability reporting and technological advancements, while maintaining a focus on audit quality and public trust.
While LLMs are powerful tools, they're not replacing human auditors anytime soon. Instead, they're likely to augment human capabilities, allowing auditors to focus on higher-level analysis and decision-making. As these models continue to evolve, they may take on more complex tasks, but human oversight and expertise will remain crucial in ensuring the accuracy and reliability of audit processes.
Understanding and leveraging LLMs can help auditors stay at the forefront of their field, improving efficiency and effectiveness in their work. As with any new technology, it's essential to approach LLMs with both excitement and caution, recognizing their potential while also being aware of their limitations.