As the potential of generative AI continues to captivate executives, today’s finance and accounting firms — tasked with sifting through a growing volume of data to provide accurate and compliant financial insight to clients — are experimenting with how the emerging technology can fit into their daily workflows.

Big Four accounting firm Deloitte, for example, recently launched an internal, GenAI-driven chatbot — known as DARTbot — which can allow its 18,000 audit and accounting professionals to ask and receive answers to complicated accounting questions in a conversational way.

“So as a professional in the field is doing research on a particular topic and in the course of a client engagement, they can use this tool to help them respond to those types of inquiries, or to assist in the audit process,” Will Bible, digital transformation and innovation leader, audit and assurance for Big Four accounting firm Deloitte said.

Rather than having to conduct a topical search and then dig deeper into accounting literature, “you can ask direct questions and get human-esque responses based on that specific literature,” Bible said in an interview together with Deloitte’s Chris Griffin, managing partner and leader of its U.S. audit and assurance transformation and technology practice.

Creating a GenAI framework

The use of AI within audit is far from new; Deloitte has been utilizing large language models for several years, Bible said. With the emergence of GenAI, the firm moved quickly to figure out how the technology could be put to best use within the field. For DARTbot, GenAI was layered over an existing accounting research tool at Deloitte, enabling faster responses from a library of knowledge audit professionals were already familiar navigating, Bible said.

 The company also tested out the DARTbot with a small group of users before rolling it out to its 18,000 audit professionals, Griffin said, ensuring that the tool was responding to queries correctly and in a way that dovetailed with their professional experiences in the field.

Deloitte also “put built in parameters within it structurally so they couldn't go off the reservation, so to speak,” Griffin said. “It would only answer questions that were germane to the accounting roadmap.”

It's also important to ensure that employees have the proper training to interact with and effectively utilize the tool; when looking to upskill one’s employees, it’s important to foster trust and reliability with the tool and to make sure “people understand that you have to apply logic and judgment to the outputs from the technology,” Griffin said. “So that's an uplift that we're all going have to go through as we learn how to use this technology in our day-to-day work.”

Keeping the focus on responsible AI

Like its fellow Big Four firms including Ernst & Young, KPMG and PricewaterhouseCoopers, Deloitte is funneling billions into new technologies,  investing $1.4 billion in a technology learning initiative aimed at boosting key skills among their employees, including those related to GenAI. However, “when you think about this technology, it's not infallible,” Griffin said. “There are concerns around security, fairness, privacy, responsibility.”

When it comes to the actual implementation of GenAI, therefore, it’s important to have a framework in place that is “transparent and explainable,” Griffin said, as well as one that fosters responsibility, fairness, and privacy.

“You’ve got to have a plan, you’ve got to think about controls, you’ve got think about the development … about, how do you make sure that it's giving you the right information?” Griffin said.

Developing “responsible AI” is a top priority for Deloitte’s fellow Big Four firms as they race to incorporate the emerging technology into their systems and processes; KPMG announced in September that it would be forming an AI and Digital Innovation Group which would center on the development of responsible AI tools, CFO Dive previously reported.

EY, meanwhile, is also keeping its focus on ethical AI solutions as it looks to bring the technology to bear in its own organization — the Big Four firm also announced an AI-driven chatbot for EY employees earlier this year, focused on answering payroll questions.   

That need for responsible AI — which can enable firms to provide a transparent trail of where their data is coming from, helping with compliance as well as other obligations — also speaks to the continued importance of the human element in the accounting and finance fields.

“So it doesn't remove people — it’s an accelerator,” Griffin said of GenAI and how the firm is incorporating the technology. “But people are still going to be incredibly important in evaluating the outputs of the technology, asserting whether or not it's coming out with a reasonable set of answers based on what you asked it.”