AI-First Accounting: Hype Or Future?

The accounting world is brimming with anticipation about how artificial intelligence (AI) will revolutionize the industry. Is this just hype, or are we truly on the verge of a categorical shift in how accounting professionals work? This editorial discusses the current status quo of AI in accounting, the technologies available today, and the challenges ahead.

AI in accounting is alluring - automation of routine tasks, enhanced accuracy, real-time insights, and the ability to process vast amounts of data very quickly. However, as with any technological leap, it’s crucial to separate the hype from reality and understand both the potential and the limitations across use-cases.

STATUS QUO OF AI IN ACCOUNTING

AI is already making an impact in the accounting sector. According to a report by Deloitte, 83% of early adopters of AI in finance and accounting have achieved returns on their investments (Deloitte, 2020). A study by Accenture revealed that companies using AI for accounting processes experienced a 70% reduction in time spent on routine tasks, allowing employees to focus on more strategic activities (Accenture, 2021). According to PwC, AI forecasting models improved financial planning and analysis accuracy by up to 40% (PwC, 2022). Early adopters are witnessing improved accuracy, efficiency, and knowledge from the use of AI in accounting.

Most common use-cases and applications:

1. Automated Data Entry and Categorization: AI-first accounting software like Jaz use AI-powered optical character recognition (OCR) tools can scan receipts, invoices and bills, extracting relevant information and categorizing transactions automatically. This saves time and also reduces human error in manual data entry.

2. Anomaly Detection: Tools such as Mindbridge use machine learning algorithms to analyze large volumes of transaction data to identify unusual patterns, which may indicate mistakes or fraud. This significantly enhances the efficiency of monthly closing processes.

3. Predictive Analytics: AI tools like Anaplan analyze historical financial data to forecast future trends, helping businesses make more informed decisions about cash flow, inventory management, and more. This can be particularly valuable for financial planning and risk management.

4. Natural Language Processing (NLP) for Regulatory Compliance: IBM Watson Regulatory Compliance and Thomson Reuters NLP systems can read through complex financial regulations and summarize key points, flagging potential compliance issues based on a company’s financial data. This helps companies stay compliant with ever-changing regulations.

5. Intelligent Financial Reporting: Tools like Workiva can assist in generating financial reports, not just by compiling data but by providing insights and highlighting key trends or issues that human accountants might overlook.

TECHNOLOGY DEEP DIVE

The AI used in accounting isn’t a singular technology, but rather a suite of tools and techniques. Machine learning, particularly deep learning, forms the backbone of many AI use-cases in accounting. For instance, the OCR systems used by AI-first accounting software for document scanning use advanced neural networks, a type of deep learning algorithm particularly effective at handwriting recognition.

Predictive analytics in accounting often uses modeling techniques like regression analysis and time series forecasting. These methods analyze historical financial data to predict future trends with increasing accuracy as they ingest and encode more data over time. More advanced systems might use ensemble methods, combining multiple models to improve prediction accuracy.

Natural Language Processing (NLP) is another crucial technology in AI-first accounting software. NLP allows software to understand human language, enabling features like quick onboarding, chatbots for customer queries, or automated report generation. Recent NLP advancements, such as transformer models like GPT, LLAMA, and BERT, have significantly improved the ability of models to understand context and nuance within the accounting and finance domain.

Robotic Process Automation (RPA) is often considered the predecessor of AI. While not “intelligent” in itself, RPA automates repetitive tasks and workflows, freeing up accountants’ time for more strategic work. When combined with AI-first accounting software, RPA can become “magical” and capable of handling more complex, judgment, or reasoning based tasks.

CHALLENGES AHEAD

Despite the current applications, the adoption of AI in accounting firms faces several challenges:

1. Data Quality & Explainability: AI requires high-quality and consistent data to function effectively. Many organizations struggle with data silos and inconsistent data formats across different systems. Many AI models have neural networks of “closed weights”, and operate as black boxes, making it difficult to explain their chain-of-thought. A study by KPMG found that 92% of C-level executives are concerned about the impact of data quality on AI (KPMG, 2019). AI-first accounting software that adopt “chain-of-thought prompting” (Wei et al., 2022) can provide better explainability.

2. Accountability: The use of AI in financial decision-making raises important questions about accountability and potential biases. For instance, if an AI system makes a decision that leads to a bookkeeping error or a financial loss, who is held responsible? The most common workarounds are to limit AI’s power to drafts or non-critical workflows, but as demand for the technology increases, additional guardrails will be required.

3. Cost of Implementation: While large accounting firms can afford to invest capital expenditure in AI technologies, smaller accounting practitioners may find the initial costs prohibitive. This could lead to a digital divide within the industry. AI-first accounting software seek to bridge this divide by offering the most important solutions to practitioners with affordable pricing and onboarding support, so that they can in turn serve their small-medium customers.

4. Knowledge Gap: A survey by the Association of International Certified Professional Accountants found that 80% of accountants believe they still need to develop their digital skills for the future (AICPA, 2019). International and national bodies like the Institute of Singapore Chartered Accountants (ISCA) organize continuous education modules, interactive workshops, and hands-on activities like “AI Demo Days” for accountants to build knowledge, experience, and familiarity with AI.

5. Cybersecurity Concerns: As accountants become more reliant on digital technologies, they may become more vulnerable to cyber-attacks. It’s critical for organizations and software to use enterprise-grade AI infrastructure (Amazon Web Services, Google Cloud Platform, Microsoft Azure).

CONCLUSION

AI-first accounting is not just hype. It’s already yielding tangible returns in large firms. For small-medium practitioners, the adoption of AI will require (i) openness to next-gen technologies like AI-first accounting software, (ii) onboarding support, and (iii) adequate team trainings.

“The King is dead. Long live the King” is a translated phrase from Old French, which reinforces the continuity of the way of life amidst a significant event. This expression reflects the inflection point of accountants today. As traditional processes give way to new, the essence of the profession remains, but the tools will transition.

---

Kenneth Yeow is Partnerships Director @ Jaz in Singapore. Jaz is an AI-first accounting software for business owners, accountants, and finance teams.