‘Human in the loop’: As companies integrate AI, governance and literacy are key

Middle market organizations are preparing to ramp up investments in artificial intelligence over the next two years, according to findings from the RSM US Middle Market Business Index Special Report: Workforce 2026. For companies to get the most mileage possible out of AI tools and capabilities, a clear governance framework and AI literacy initiatives will be essential guardrails in an evolving space, RSM professionals say.

Out of 405 respondents surveyed in the fourth quarter of 2025 for the report:

• 38% said they plan to significantly increase spending on AI technology to enhance worker productivity over the next two years.
• 37% plan to somewhat increase such spending.

Among executives who expect staffing over the next year to be at least somewhat challenging, 77% said they are either preparing for or investing in AI.

John Huyette, an RSM principal and the firm’s national leader for AI risk consulting services, is seeing more companies baking in AI governance frameworks early on in their adoption of AI. That makes governance priorities even more essential in a competitive market.

While a typical governance framework addresses values, regulatory compliance, security and privacy, AI governance frameworks involve additional considerations for bias, ethical use, consistency, explainability and accountability.

“Governance is not meant to put the brakes on AI, but rather to guide good decisions,” he says. “As more companies progress with AI, they need to understand the quality assurance they provide, how they monitor that, and other human-in-the-loop considerations.”

In addition to governance, companies need to be aware of risks AI might introduce from an identity access management perspective, Huyette says. If an organization develops an AI agent that uses data from one function of the business, for instance, could that data end up in front of someone who shouldn’t have access to it? Businesses need to develop policies to mitigate such data leakage, he says.

An AI governance and strategy risk assessment can help organizations identify gaps in how they incorporate the evolving technology across operations.

AI literacy training

Once governance processes are in place and employees begin testing and getting comfortable with AI tools, literacy training around the technology can help alleviate pain points that arise.

“AI literacy training helps people understand how AI applies to them and can solve problems to augment their role, not replace them,” says Robbie Beyer, national leader of RSM’s data science and AI practice for data and digital services.

Aspects of that training might involve:

• Understanding how best to prompt an AI tool and refine those prompts
• How to determine responsible uses of AI across different functions of the business
• What type of information should not be used with AI tools because of data privacy or other sensitivities

Still, some adoption challenges will likely arise along the way. In a market where many roles are already complex, adding AI skills to the mix brings another layer, says Ana Minter, an RSM principal who specializes in process improvement, automation and AI.

“There's sort of this pinch in skill sets that are available as we all learn together,” she says. “It's like teaching everyone a new language, a new way to think, all at the same time.”

In figuring out how to translate the opportunity to business growth, she adds that companies should not leave people out of the AI equation but rather equip them with the right tools. “The only way you're going to stay in that race is to prepare your people.”

Companies thinking about how to implement AI tools should start with smaller tasks and scale up from there, says Tuan Nguyen, an RSM economist. That’s usually the clearest way to measure the impact on productivity.

“Especially for small and middle market firms, it’s going to be a challenge in terms of budget to invest in AI on a larger scale,” he says. “Smaller-scale projects should be the top priority that organizations can then study the impact of.”