Rethinking audit—Part 3: How will audit delivery change

The series will examine:

1. Who values audit?
2. Why is audit important?
3. How will audit delivery change?
4. What value can audit bring?

Part 3: How will audits be delivered in the mid to long term?

If I roll the clock back some twenty years or so, we audited without the help of computers. In my training days, management held records of all the transactions of the business on huge dot matrix ledger printouts. We would sit with rulers and calculators and check the mathematical accuracy before transposing sub-account balances on to seven column paper, validating this against the information prepared by the entity we were auditing.

The introduction of computers initially saw us replicating exactly what we did manually onto the computer. We created documents in Word and Excel that looked exactly like seven column paper. Then electronic audit files started to develop, followed by systems to incorporate methodologies, audit tests that should be conducted and integrated access to auditing and accounting standards. Alongside this, “Computer Aided Audit Techniques” or CAATS started to develop. This is where the power of technology really started to add value, rather than simply using a computer as a filing system. With CAATS, we could work with IT contacts at the audited entity to download entire populations of data. The CAATS could quickly look across all the data and spit out anomalies. This could then focus the follow up testing conducted by the auditor—was there a problem with the system? Did someone input something incorrectly? Or was something else going on?

Nowadays, audit systems are incredibly sophisticated, and larger audit providers often have their own in-house developed system. They have integrated guidance and tailoring for the industry, size, and nature of the business. They include analytics tools and often seamless data extraction from the audited entity. Analytics (the evolution of CAATS) will not only look at data from the entity but may well overlay external data such as market pricing, weather, geography—whatever is pertinent to hone in on areas of risk, both for the business and for the material correctness of the financial reporting. Auditors can even use drones to survey client assets in remote locations. The use of technology is limited only by the imagination of the auditor.

If this is the present, then what of the future?

Imagining a tech-first audit

The tendency is to think of using technology to replace procedures and processes that are already carried out by people. Arguably the best leading edge “audit tech” is still a very sophisticated filing system with some built-in analytics and increasingly integrated AI. These generative and agentic AI tools are being used to speed up or take over human work, perhaps by summarising audit working papers, enhancing research into relevant audit and accounting standards, or automating some of the cross-checking of audited balances to the financial statements.

Yet this is simply incorporation of current tech to replace current process. I would like to take this opportunity to throw down the gauntlet and challenge the “how” of performing audits.

To quote Richard Susskind in his excellent book How to think about AI: A guide for the perplexed, professionals may take on a “’not us” mentality when considering the work they do and the potential to apply AI to it:

”When Daniel Susskind and I were interviewing a wide range of professional workers in preparation of our book The Future of the Professions (2015). Time and again, we noted that professionals see much greater scope for AI in disciplines other than their own. […] Many, if not most, professional workers regard themselves as artists in their own craft. They look upon their work as the very embodiment of what machines will never be able to do. All manner of biases and dissonances are going on here, but the undeniable fact is that the professions and the white-collar workforce are overflowing with people who believe that AI has massive potential but ‘not for us.’”

From my own experiences, working within an audit business, a view commonly expressed was that technology would only take us so far, but the realms of professional scepticism and judgement would remain exclusively human.

I would like to challenge that presumption. I suggest that judgement is actually a form of pattern recognition. In an audit context, judgement may come from a very experienced professional or group of professionals comparing a set of numbers, together with understanding of the business and the market to similar circumstances and patterns experienced in the past. Technology is already very good at this kind of task, albeit sometimes with significant flaws. It is rapidly getting better at it. One example is around medical diagnoses. Large language models (LLMs) can hold vastly more cases as a readily accessible data bank than any individual would experience in a lifetime’s work, and they are being trained on pattern recognition and medical application. It is, however, interesting to note that increased usage of AI may lead to de-skilling of the human practitioner.

If AI in its current state is being developed to solve complex medical diagnostic challenges, then I suggest that it will also be able to provide the judgement element of an audit.

I also mentioned professional scepticism as being a distinctly human attribute. What I mean by this is the ability to interrogate the story behind the numbers, rather than accept an explanation at face value. The numbers in a set of financial statements tell the story of what has been happening in that business over the period they relate to. When we audit those numbers we ask “why?” a lot. Why have sales grown or dipped, say? And then, does the explanation really stack up, ideally against some form or independent evidence? It is hard to see how an entirely AI-based audit would replicate this process. But would it need to?

The challenge here is not to replicate how humans audit, but to get to outstanding audit quality using the power and capabilities of computing. Large Language Models (LLMs) using huge repositories of data could assess the transactional and reported numbers against external trends and data. For example, interest rates, pricing of raw materials and finished goods, data on weather, costs of capital, and more. Rather than applying professional scepticism, AI can use sheer processing power and massive data sets to stress test, look for patterns, and identify anomalies and areas of risk or error.

Given this, the methodology of audits could be very different in the future. To quote Richard Susskind again:

“Our systems and machines are becoming increasingly capable; the underpinning technologies are advancing at an explosive, exponential rate; there is no apparent finishing line; we haven’t invented most of the technologies that will change our future lives.”

The audit profession should be thinking further ahead than the deployment of technologies readily available today. It should be designing for a future where the whole methodology, the underpinning systems and processes, and indeed the role of the human within the audit are re-imagined. In our present we are seeing GenAI at work in its early days. The future is likely to bring massively capable technology able to operate at least at the level of human experts.

Considerations in designing the future of audit

As an innovator and system designer, one technique I often use when working with teams trying to imagine a different future is, “what would Google do?”

Google has all the tech power and capability needed to build a seamless end-to-end process. It also has the minds and innovative culture to design how things could be done differently, starting with the outcome they want from an audit. As a new entrant to the market, it would be unencumbered by historic processes, organisational cultures and traditional ways of working. If Google wanted to enter the audit profession, how would they do it, and what could we take from that mindset?

On top of the rapidly accelerating landscape of AI advancement, we are seeing increasing amounts of private equity coming into the audit profession. Perhaps the rapid acceleration of technological opportunity is drawing their interest? The potential for increased efficiency and enhanced quality delivered through system and process redesign could be part of the opportunity. This underlines the need for auditors to keep up with the speed of change in the wider world.

What’s next for audit firms?

So, what’s next for audit firms? Think through the value of the human auditors. One senior audit partner I interviewed said that as long as clients have people, audits must be led by people. This implies that it is the inter-personal role that is critical. How do we focus on training auditors of the future with the ability to turn tech-driven outcomes into a story that resonates on a human-to-human level? Or leaders who can translate the audit technology output into insight that executives and boards can understand, trust, and act upon?  Articulating the value of the human auditor will be critical for the profession to rise on the tide of technological progress.

About Jo

Jo Rhoden is a Fellow of the Institute of Chartered Accountants in England and Wales. She began her career in London in audit and consulting before moving to Sydney, where she focused on fast-growing technology and biotechnology companies, including audit and due diligence for venture capital firms. Moving to a Big Four firm, she helped transform the audit process by integrating innovation, data, digital, and design techniques, building a team of “audit designers” to enhance client experiences and later leading this work globally through a client experience programme. Jo also led brand strategy and developed a Global Audit Committee programme addressing major market-facing issues. She currently serves as Chair of the Audit and Risk Assurance Committee at the University of Bath and works as an independent consultant in strategy, design, and innovation, with a focus on the future of professional services.

Source’s audit research

Learn more about Source’s research into the future of audit here, or contact us.