Amid AI uncertainty, good data is the north star, says RSM’s digital leader

Keeping pace with changes in artificial intelligence is not for the faint of heart or wallet, particularly if you’re a middle market business that requires prudent resource allocation, says Sergio de la Fe, RSM enterprise digital leader.

Midsize businesses require a clear understanding of their data and foolproof processes to keep that information current to stack the deck in their favor.

“AI has accelerated the use of good data,” says de la Fe, a partner and former chairman of the RSM board whose current role includes oversight of AI development on behalf of the firm. “Organizations that don’t have good data are not going to be able to apply AI. Fixing your data is important.”

Like most companies, midsize businesses are keen on leveraging generative AI technology; RSM's Middle Market AI Survey 2024: U.S. and Canada shows that 78% are either formally or informally using platforms such as Microsoft Copilot and ChatGPT. But more than half (54%) say AI has been harder to implement than expected, and a solid majority (67%) agree they need outside help to get the most out of these solutions.

De la Fe believes one reason for the difficulty midsize businesses face is the breakneck pace of AI development—which leaves many companies puzzled about where to spend their money.

“I think many companies don’t know where to start,” he says, adding that a big question is “build or buy.” Deploying resources to build an internal AI process, only to find out that a software vendor will supplant those efforts by doing the same thing with more resources and specialized expertise, can be daunting, he adds.

In the recent RSM Middle Market AI Survey 2024: U.S. and Canada:

RSM, which faces similar challenges, currently has about a dozen use cases for AI in various stages of implementation, de la Fe says. Chief among them are ways to reduce the time spent on document research and writing.  

“We want to drive efficiencies in simple mundane tasks,” De la Fe says. That goal is aligned with findings in the RSM Middle Market AI Survey 2024: U.S. and Canada, which showed that the leading expectations for AI technology were improved quality control (58% of respondents), enhanced customer service (51%), automation of repetitive tasks (45%) and increased productivity and creativity (45%).

Consider a pilot in RSM’s Washington National Tax practice that uses an AI large language model platform to develop drafts of custom position papers for clients: the technology marries information from RSM’s historical reports with a library of tax law to give practitioners a leg up, delivering a preliminary document in seconds when it used to take hours.

“You get that first draft much faster,” says de la Fe. “You can ask more quality questions to refine the position. That is a Holy Grail.”

In the firm’s consulting practice, RSM is deploying similar technology to determine whether clients have the necessary controls in place to meet myriad compliance standards, mapping complex regulatory requirements to the processes and controls at a particular business and vice versa.

“It also gives you improvements,” says de la Fe. “That is massive.”

One misplaced fear about AI technology is that it will lead to reduced quality. But for AI solutions to be effective, de la Fe says they always require the human touch, adding that successful use cases eliminate much of the front-end grunt work, allowing professionals to spend their time on thoughtful analysis.

“That speed to quality enhancement is game-changing,” says de la Fe.