AI Business · · 4 min read
What Happens to Human Auditors If KPMG's AI Asks Better Questions?
KPMG's AI trial suggests a future where machines outperform humans in critical audit tasks. We explore what this means for the industry and the role of human expertise.

Earlier this year, KPMG deployed over 100 AI agents. They scrutinised the financial statements of a UK-listed FTSE 250 company.
This wasn't hypothetical. The AI ran alongside a conventional audit by KPMG's human accountants. It examined the same books, collected and analysed data, and even drew conclusions.
The Financial Times reported striking findings. Cath Burnet, KPMG's UK head of audit, said the AI went "so much deeper." It also asked "better questions" than human counterparts in several cases.
Imagine that: machines outperforming seasoned professionals in the very act of financial interrogation.
On the surface, it sounds like a technological triumph for KPMG's "Project Exponential." But that's not the interesting part.
The real story goes beyond AI's capabilities. It's about profound implications for an industry built on human judgment, scrutiny, and trust.
What does it mean when the machines start asking better questions than we do?
That raises three questions.
Question 1
What Did KPMG's Project Exponential Actually Reveal About AI's Audit Capabilities?
Project Exponential was an ambitious trial. Over 100 AI agents joined a parallel audit process. These agents were built in just four months, showing rapid AI development.
They didn't just assist with small tasks. AI agents worked across different audit stages. They collected, analysed, and recorded data. Then they drew conclusions and draft assessments.
The experiment showed AI could investigate financial issues more deeply. It posed more incisive questions than human auditors, KPMG reported. This implies a significant leap in AI's analytical power within a complex domain.
For context, traditional audits are labor-intensive. Large teams of junior accountants sift through vast data under senior supervision. The AI system mimicked, and sometimes surpassed, this human effort. An audit partner signed off the experimental audit. However, it was not used for the formal sign-off of the company's accounts.
This trial showed AI handles sophisticated analytical tasks. It goes beyond data crunching to actual inquiry and preliminary judgment. And that brings us to the second question.
Question 2
How Do AI Agents Manage to "Go Deeper" and Ask "Better Questions" than Human Auditors?
Well, consider the sheer volume of data in a large company's financial statements. Human auditors, even skilled ones, face limits. These include time, cognitive load, and potential for error or oversight.
AI agents, however, process vast datasets at incredible speeds. They cross-reference every transaction and line item. This is done against databases of regulations, policies, and historical data. They work with relentless consistency.
Imagine a diligent kirana store owner. They manually check every item on a large supplier invoice against stock records and pricing agreements. It's tedious, time-consuming, and prone to mistakes.
Now, picture that owner using a digital system. It instantly flags any discrepancy, no matter how small. It even suggests follow-up questions. This is based on patterns learned from thousands of invoices. That's how AI agents operate.
AI agents excel at pattern recognition, anomaly detection, and data synthesis. KPMG's Burnet noted "full visibility and full traceability" of the AI's work. It recorded interactions down to individual characters and timestamps. This highlights the AI's methodical and transparent process.
This allows AI to identify subtle inconsistencies or unusual patterns. A human might ask a basic question. But the AI, having processed more data, can formulate a deeper, more targeted inquiry. What does this mean for human judgment? However, human intervention was still needed. An AI agent failed to match regulations with company documents. This shows AI's strength is in processing, not nuanced interpretation or ambiguity. And that brings us to the final question.
Question 3
What Does KPMG's AI Audit Trial Mean for the Future of Human Auditors and the Industry?
Efficiency is the most immediate implication. KPMG saw "clear time-saving opportunities" with AI-based audits. Using AI should mean faster and potentially cheaper audits. This compares to relying solely on human accountants.
For business leaders, this could mean more timely audit reports. It also suggests potentially reduced costs. KPMG did not say if these efficiencies would lower audit fees, however.
The human auditor's role will evolve. No more hours on labor-intensive data collection and initial analysis. Human auditors might shift to higher-value tasks. These include interpreting complex AI findings. They will exercise professional judgment in ambiguous situations. They will also handle client communication and negotiation. How will the human auditor's role evolve?
KPMG plans to expand this project to up to 20 large-scale audits. It targets providing AI-based audit services long-term. This signals a serious commitment to integrating AI into its core offerings.
However, significant concerns exist. A report by the Centre for Public Interest Audit found 23 percent of company audit committee chairs had "limited" confidence in AI's use. The UK's Financial Reporting Council (FRC) also warned AI could pose a "clear threat to the integrity and independence of auditors."
Risks include over-reliance on AI, weakening professional skepticism, data security breaches, and AI generating false information. The future likely involves a hybrid model. AI will augment, rather than fully replace, human expertise.
The Bigger Picture
KPMG's Project Exponential is more than an internal experiment. It's a powerful signal to every business leader. AI isn't just automating tasks. It's showing superior analytical capabilities in uniquely human domains.
For businesses, the question isn't if AI impacts operations. It's how fast it redefines valued skills and processes. An AI asking "better questions" is a wake-up call.
We must rethink human roles. Focus on oversight, ethical judgment, and complex, creative problem-solving. Machines still struggle with these areas.
The future of work isn't humans versus machines. It's humans intelligently leveraging machines. This pushes the boundaries of what's possible. The audit world is just the latest to learn this lesson.
Quick Answers
What did KPMG's AI audit trial reveal?
KPMG's trial, called Project Exponential, found that AI agents asked 'better questions' and investigated financial issues 'deeper' than human accountants during an experimental audit of a UK-listed company. The AI was able to collect, analyse, and record data, and draw conclusions.
How can AI agents ask better questions than human auditors?
AI agents can process and cross-reference vast amounts of financial data against regulations and historical information at high speed. This allows them to identify subtle patterns and anomalies that might lead to more targeted and incisive inquiries than a human auditor could typically formulate.
What does this mean for the future of human auditors?
The trial suggests human auditors' roles will evolve, shifting from labor-intensive data analysis to higher-value tasks like interpreting AI findings, exercising professional judgment in complex situations, and managing client relationships. AI is expected to augment, rather than fully replace, human expertise in auditing.
Sources
- Business Standard18 Nov · 12:58 IST
- Financial Times18 Nov · 10:30 IST
Facts are credited to the outlets above. The analysis, calculations and Signal Check are by the Finnect AI Insights desk.
