While generative AI is still a relatively novel technology, the Big Four see vast potential to transform their service offerings and create a few new ones along the way, and so are working hard to build it into as many of their existing services as possible. It would seem there are few areas that technology leaders think can't benefit from generative AI, and so are looking for it to enhance their current capabilities.
For instance, Will Bible, Deloitte's audit and assurance digital transformation and innovation leader, felt there was great potential to enhance the firm's audit servings with generative AI. Rather than talk about entirely new ways to conduct an audit, he spoke in terms of how AI can augment the capacities of a human auditor in what they do today. He noted that, regardless of AI advances, people will still demand independent accountability for the entities using the AIs, which means audit and assurance engagements.
"I think people will always need audits because what key stakeholders look for is third party verification, reasonable assurance levels for financial statements. The question, then, is how that product evolves. I think, to address how different that product might be, we want to think of the attributes of the product and what people want, which is more and timely financial information reporting from companies, so the big question in my mind is can we offer more frequently audits, can we offer more audits of information, can that product evolve in the back of improved technology?" he said.
Audits of AI themselves is one possible way the audit and assurance practice might evolve. Bible said the firm is currently devoting R&D resources to the question of how best to conduct such engagements, especially with regard to what he called the "black box problem" where "by definition you can't get in and assess the inner workings."
"So what is the business process that the technology has been applied to, and what are the outcomes, and what are you measuring against? Because even though you have a black box, you can still measure the impact or outcomes from that black box. So, I do see an opportunity to provide assurance or verification of those outcomes, or an assessment of those outcomes," he said.
Chris Griffin, managing partner of transformation and technology at Deloitte Audit and Assurance, noted that generative AI will likely also be incorporated into ESG frameworks, and so there will likely be demand from clients for assistance on either how to manage their own AIs in the context of ESG frameworks, or how to assess the AIs of other organizations. This is part of what he believes will be an overall expansion and enhancement to the kind of assurance they will be able to offer.
"We want to see how this can be a facilitator for these market demands for different types of assurance, which we can provide," he said.
This is but one part of many areas where Deloitte plans to implement AI, with Bible saying they're taking a "multi-disciplinary approach."
Paul Goodhew, E&Y's global assurance innovation and emerging technology leader, also said that generative AI enhancements are planned throughout the firm. He described E&Y's overall approach towards AI as "service line specific," meaning that there are specific considerations pertaining to clients that contextualize its AI moves. For instance, he said the firm invested $1 billion to upgrade its assurance technology overall, which included a generative AI component to enhance the capabilities of professionals in the field. Regardless of the specific service line, though, E&Y sees great potential for generative AI within it, at least partially due to the commonalities of what it can offer.
"Where the AI supports our professionals with predicting, with the use of data analytics and identifying patterns and anomalies, where AI can provide an interactive assistant for our professionals, we ultimately see a number of big opportunities to the full range of AI use cases where we can start to introduce more and more of our technology to our people… Given the amount of processing power and sophistication, the launch of generative AI opens up more potential for how we can support our people and transform how they work," he said.
Cliff Justice, KPMG U.S. enterprise innovation leader, said one of the challenges of embarking on its AI journey has been sorting through what's real and what's hype. A great deal of fog has emerged in the field, obscuring what AI can and cannot do. In KPMG's own AI incubator (each of the firms alluded to having something along these lines), much time is spent on experimenting with potential new services and enhancements to existing one, but also on examining whether the underlying technology even works as advertised.
"As you might imagine, there are hundreds of tools now that claim to be AI and so there's a lot of sorting through the smoke to see which tools are really generative AI. It's different than the machine learning we've been working with for five, six years. It's really a different type of AI… so we have to experiment to see if it works in our environment, test it, and see if we can put it into production with live engagements," he said.
As for how they will be getting these AIs, right now the firms are generally partnering with tech companies like Microsoft or OpenAI to access a foundational model that works as a base for heavy customizations and modifications, often through direct training on firm data, or they are buying or licensing models directly. This is only the beginning, though, and each of the firms envisioned a future where they are not only a savvy consumer of AI technology but a producer as well.
Joe Atkinson, chief products and technology officer with PwC, said that–given the firm's huge amount of resources and expertise–it makes sense for them to eventually develop their own proprietary solutions.
"We're clearly consumers of AI technology today from our partners, but not everyone has the investment capacity we enjoy as a large company. We have the benefits of scale and a talent portfolio most would love to have. Those combined assets give us the ability to build our AI factory where we take our use cases and [see] how we can improve our own business and clients' businesses and deliver [value] in different ways," he said.
E&Y's Goodhew said they were on a similar track, and noted that they already have 11 AI-powered assurance solutions of their own in the pipeline, soon to be released. Reflecting the firm's service line-specific approach, E&Y has already released a chatbot available for clients that can answer simple HR and payroll questions.
Beyond enhancing their own capacities with AI, the firms also see potential in doing the same for their clients. Deloitte's Bible said they view this sort of assistance as essential to their overall AI goals, as it is highly likely that they will need help with their own systems.
"First and foremost is [our goal] of trying to help clients as they navigate adopting these technologies into their businesses and, in particular in audit and assurance, where we're the auditors of financial statements, and making sure we consider that in our audit approach and design as people adopt this technology into accounting and financial reporting," he said.
Getting a little more specific, PwC's Atkinson said he could see a future where clients turn to PwC to help them manage their AIs as they currently handle other tasks on their behalf.
"I can envision in the future a kind of 'AI as a service' model, like a managed service environment as we start getting more into managed services, like operationally managing the environment for our clients in applying AI. … I think we will provide more and more of our clients those options as we continue to accrue capabilities," he said.
However, the firms are not necessarily driving this effort. Leaders agreed that it is the market that makes these sorts of investments necessary. If an accounting firm can't understand their client's technology, its professionals will be hard pressed to provide adequate service. But at the same time, it is also about keeping up with changes in the entire professional services world, not just in the accounting world but overall, according to KPMG's Justice. While he was unsure what the world of professional services would look like specifically in five years, he was confident it would involve AI.
"Against the markets we're in, and against the client demands coming our way… In the next five years we'll see major changes in professional services across the board. The world we're in–legal, medicine, all types of knowledge services–are going to rely more and more on AI," he said.
