The Singapore office of this Big Four firm audited over 280 listed companies and large private enterprises annually, with audit teams ranging from 5 to 40 professionals per engagement. Traditional audit methodology relied on statistical sampling — testing 25 to 60 transactions from populations of thousands or millions to form an opinion on account balances. While statistically valid, this approach inherently accepted a degree of detection risk.
Audit hours were climbing 8% year-over-year as regulatory expectations increased, client businesses grew more complex, and accounting standards became more judgment-intensive. The firm's audit partners estimated that 40% of total audit hours were spent on routine transaction testing and data gathering — work that added limited value to the audit opinion but was necessary under current methodology. Senior associates were spending three to four days per engagement just obtaining, formatting, and reconciling client data before substantive testing could begin.
Competitive pressure from other firms deploying AI-augmented audit technologies was a growing concern. Two major audit clients had specifically asked during tender renewals whether the firm used AI to enhance audit quality. The firm's global network had developed AI audit tools, but these were designed for Western accounting systems and did not integrate with the diverse ERP platforms, accounting software, and data formats used by Singapore and Southeast Asian clients.
Pertama Partners' AI Readiness Audit assessed the firm's existing audit workflow, data extraction processes, and technology infrastructure. We analyzed how audit teams currently handled data from the 15 most common accounting systems used by Singapore-listed companies and identified that data extraction and formatting alone consumed 18% of total audit hours — a massive automation opportunity.
Our AI Pilot Program developed a suite of AI-powered audit tools deployed across 12 pilot engagements. The data ingestion module automatically extracted, standardized, and reconciled financial data from diverse client systems. The full-population testing engine analyzed 100% of transactions using anomaly detection algorithms that identified unusual patterns, potential misstatements, and areas requiring auditor judgment — replacing the traditional sampling approach with comprehensive coverage. A journal entry testing AI examined every manual journal entry posted during the period, flagging entries meeting defined risk criteria with explanations of why each was flagged.
Executive Training with the firm's leadership addressed the professional and regulatory implications of AI in audit. We worked with the firm's methodology team to ensure AI-assisted procedures were documented in a manner acceptable to ACRA and aligned with Singapore Standards on Auditing. Team Training prepared audit teams to interpret AI-generated findings, apply professional skepticism to AI recommendations, and document AI-assisted procedures in workpapers.
"Sampling has been the foundation of audit methodology for decades. Pertama Partners showed us that AI lets us move from reasonable assurance based on samples to comprehensive assurance based on complete data — and our clients and regulators see the difference."— Kenneth Loh, Head of Audit & Assurance
Every transformation starts with a conversation. Let's discuss your challenges and opportunities.
Discuss Your Challenge