Within the middle market, the use of AI is the rule, not the exception.

An overwhelming 91% of total respondents said their organization uses AI, either formally or informally, in business practices. This marks a significant increase from the RSM Middle Market AI Survey 2024, when 78% of overall respondents reported using it.

The numbers are similar when respondents were asked specifically about their use of generative AI, with 91% of total respondents saying their organization uses it, a noticeable increase from the 77% of overall respondents who said the same thing last year. 

“Companies recognize that AI is not a fad, and it's not a trend,” says Joseph Fontanazza, risk consulting AI governance leader for RSM US LLP. “Artificial intelligence is here, and it's going to change the way everyone operates, the way things work in the world. Companies don't want to be left behind.”

And organizations are doing more than just dabbling in AI. They are devoting energy to planning and developing the best ways to use the technology, evidenced by the 79% of respondents who reported having a strategy or roadmap to adopt AI. This consisted of 37% who said they have a well-defined strategy, and 42% who said they are in the early stages of development.

Companies are also striving to integrate AI into their systems. One-quarter (25%) of respondents who know whether their organization uses generative AI tools said it is fully integrated across their organization’s core operations or workflows. For an additional 43%, it is integrated across some of their operations and workflows, a significant increase from the 34% of respondents who said this in the previous survey.

These results indicate the desire of many companies to leverage AI for enhanced operational efficiency and innovation.

“Many organizations see the potential for the technology and want to implement it right away,” says Robbie Beyer, director of data science and AI for RSM US. “Others are waiting, then jumping in to execute where they think the real value is. But a few are sitting on the sidelines and feel like AI is too risky. They’re not seeing what's possible, and that might be a problem down the line for them.”

When it comes to the specific AI tools that companies prefer, ChatGPT remains the most-popular generative AI platform. Three-quarters (75%) of respondents who use generative AI tools reported using the application, which was a slight decrease from the 82% of respondents who said they used ChatGPT in the previous survey. This decrease may indicate that companies are diversifying their AI tools.

The second most-popular generative AI platform was Microsoft Copilot, which 43% of respondents said they are using, an adoption rate that has grown considerably from the 30% of respondents who said they used the app in the last survey.

Google Gemini came in as the third most-popular generative AI platform, with 35% of respondents saying they used the application. This is up from the 22% of respondents who said they used Google Bard (the predecessor of Gemini) in last year’s survey.

AI has tremendous potential to improve productivity and efficiency. One reason for this is because AI can free up employees to focus on tasks that require human ingenuity.

“Some of the basic functions that AI has helped automate—note taking, report development, putting together decks, creating proposals—have been significantly streamlined,” says Sonya King, director and AI leader of management consulting for RSM Canada. “High-performing individuals can hit the ground running, quickly take on important tasks that cannot be automated and use their critical thinking skills to develop creative solutions.”

The most popular use of AI was for text generation and summarization. Almost half (49%) of respondents used it for this purpose. Workflow development came in second among respondents, with 45% of them using AI for this task. Intelligent forecasting/demand planning tied with sales/marketing content and communications for the third-most popular use cases at 40% each.

“You’re going to see more creative ideas because people won't be bogged down with monotonous tasks,” says Matt Franko, principal in RSM US’s cyber practice. “To help them along, companies need to know what AI solutions are available within the tool sets they currently have, and they need to find out what opportunities exist in other tool sets.”

In the quest for innovation and efficiency, companies have identified numerous functions they believe AI can enhance. Of the organizations that use generative AI, 58% employ it in data analytics, making this the most popular function. Close behind are the 57% of respondents who use AI in their IT function, and the 48% who use it for customer service.

Not surprisingly, these are also the areas where respondents reported the greatest time savings from AI adoption. Half (50%) of organizations said the technology has saved them time with IT projects. AI also saved time on applications such as data analytics (45%) and customer service (39%).

Among organizations that use generative AI, many functions saw a statistically significant increase in AI usage compared to the 2024 survey. These include marketing/communications (39% vs. 30%), finance/accounting (29% vs. 22%) and supply chain management (21% vs. 15%).

Furthermore, the RSM US Middle Market Business Index Special Report: Cybersecurity 2025 reveals that AI implementation has rapidly evolved into an imperative for organizations that want to keep their data safe. In fact, an AI arms race has developed between businesses and bad actors who want to harness the power of AI to launch sophisticated attacks. 

“Cybersecurity is a great example of how AI can be extremely valuable,” says David Brassor, managing director of technology advisory services for RSM Canada. “The ability to sift through all the noise—figure out which cyber risks are relevant to the organization—and deliver actionable findings and recommendations is really where AI is providing significant value.”

Challenges remain

In general, AI performs impressively when executing a narrow task. But AI’s effectiveness diminishes if it is asked to handle multiple functions simultaneously. True success with AI requires a deliberate, targeted approach.

“Some companies know what they’re trying to accomplish with AI, but others are at square one in terms of their readiness,” Beyer says. “Different organizations are in different places. Companies often need help putting together an implementation strategy and roadmap that will guide them toward what’s possible.”

The survey confirms Beyer’s point that companies are not always ready to implement AI into their organizations. Over half (53%) of organizations that adopted and implemented generative AI believe they were only somewhat prepared to do so. In addition, a combined 10% believed they were either not very prepared (9%) or not at all prepared (1%).

The top reasons respondents cited for this lack of preparedness were a lack of in-house expertise (39%), a lack of clear AI strategy (34%) and data quality challenges (32%).

“Some organizations wanted to operate in the early-adopter space, with the approach of move fast and break things, then figure out the details later,” Fontanazza says. “But the first and most important thing the company needs to identify is what their AI strategy looks like. What direction are they going in? Do they want to be on the leading edge, or are they more conservative in their approach? The strategy is going to dictate the whole process.”

Even among respondents who considered themselves prepared for AI, it was not always a smooth  endeavor. When it came to implementation, almost all respondents (92%) said they experienced challenges.

For these respondents, concerns about data quality was the top issue (41%), followed by worries about data privacy and security (39%) and insufficient internal skills/expertise (35%).

“There is a blast radius when turning on AI,” says Diego Rosenfeld, principal at RSM US. “The effects are much bigger. Sensitive data can be leaked, or unforeseen consequences can come up. Organizations need to strike the right balance between allowing employees to experiment with different use cases and being responsible and efficient.”

Companies are acknowledging the challenges of integrating AI into their organizations. A strong majority (70%) of respondents said they needed outside help to get the most out of their AI solutions.

Furthermore, 62% of respondents said generative AI has been harder to implement than they expected, a significant increase from the 2024 survey (54%).

Despite these issues, companies were generally optimistic about AI’s impact on their organization. An overwhelming majority (88%) of respondents said it has affected their organization more positively than expected. This is substantially more than the 31% of respondents who said they experienced negative or unexpected consequences.

The most common negative consequence was workforce-related impacts—such as job displacement, skills gaps or resistance to change—which 39% of these respondents identified as a problem. A slightly smaller percentage (36%) named data quality issues—such as incomplete, inaccurate or biased datasets—as a challenge. One-third (33%) identified operational inefficiencies—such as integration challenges and system failures—as an issue.

“Having an airtight AI governance policy that is well communicated and well adopted throughout the organization is absolutely key to success,” King says. “Firms need to monitor their AI governance policies. They must make sure that no one is implementing AI without the full visibility of a governance board or committee. Companies really need a centralized function that controls AI solutions and ensures they align with an agreed-upon policy.”

Within the middle market, investment in AI remains strong. Over three-quarters (76%) of respondents said they have a budget for making investments in AI. Of these respondents, a combined 88% expected their AI budget to increase in the coming fiscal year (31% expected a substantial increase, and 57% expected somewhat of an increase). 

In addition, almost half (47%) of respondents who have a dedicated generative AI budget said they allocate funds specifically for AI outsourced consulting services.

As with any technology, AI requires people to implement and run it effectively. The human element is the most crucial factor when it comes to the successful use of AI.

“Companies can’t rely entirely on AI to make decisions or operate independently without human oversight,” Fontanazza says. “Accountability for an AI model doesn't end at implementation. People are accountable for the intake process, the design, continuous monitoring and validation. Someone needs to make the decision to take the AI offline if it's missing its mark or not operating the way it was intended. And that’s throughout the AI’s lifecycle. If companies don’t do that, they’re going to find themselves in trouble.”

For the organizations using generative AI, the most common method for preparing employees to use AI is by providing education on the safe, ethical and responsible use of AI. Almost half (45%) of respondents said they are taking this approach.

The next most popular strategy was offering workshops, certifications or courses focused on AI technologies, which 42% of companies said they are providing. Following this approach are the 41% of respondents who said they are introducing AI literacy programs for employees who do not have technical backgrounds.

“The skills gap among employees is closing because people are getting more familiar with AI,” Brassor says. “They're getting training, learning best practices and actively working on AI applications. So that’s good news for companies.”

Of course, organizations are looking for a return on their investment. Almost every respondent (97%) said they are monitoring generative AI’s success within their organization.

The top method for measuring AI’s ROI was assessing employee and team impact by analyzing concepts such as employee productivity and satisfaction. Over half (51%) of respondents said they are looking at these variables. Slightly fewer respondents (50%) said they will assess AI’s usefulness by gauging operational efficiency, while 44% of organizations said they will analyze business impact—such as revenue growth—when measuring AI’s effectiveness.

In the United States, AI regulation is fragmented, with state or sector-specific laws and guidelines providing direction rather than a single overarching framework.

Canada is developing a regulatory framework for AI, but comprehensive nationwide legislation has not been enacted.

The European Union has more comprehensive AI regulations, such as its Artificial Intelligence Act, while other nations vary in their degree of oversight.

“There are concerns about privacy and security with AI,” King says. “Because of the concept of ‘garbage in, garbage out,’ if your data is not structured and clean, or if your governance is flawed, an AI solution is not going to help you. Companies need to be aware of the problems that can develop.”

Regardless of where they are located or do business, organizations indicated they want to remain in compliance with any regulations or laws that might affect their use of AI.

The most common way that organizations are staying updated on regulations is by following industry-specific publications or thought leaders for regulatory news. Forty percent of respondents said they are taking this approach.

In addition, 36% of respondents said they remain up to date by attending conferences, webinars or training programs focused on AI legislation. An equal percentage (36%) of respondents said they are establishing internal teams to review and share updates on AI regulatory developments.

To prepare for compliance with emerging AI regulations, respondents are taking a variety of approaches. The most popular tactic is to train staff members on compliance requirements, which 45% of companies said they are doing. Upgrading technology to meet compliance standards (44%) and developing ethical AI guidelines aligned with regulations (43%) were respondents’ next most popular strategies for achieving compliance.

No one can truly predict the next stage of AI. Even by the fast-paced standards of technological advancement, AI is progressing and evolving at a furious rate. We now live in an AI-enhanced world.

A next step in the AI revolution is agentic technology, which is AI that is designed to act autonomously, making decisions and taking actions to achieve specific goals without constant human intervention. In the middle market, agentic AI adoption is relatively low, but the technology’s potential is high. Among respondents who use generative AI, 9% use Salesforce Einstein, and 5% use ServiceNow Now Assist. Both applications are examples of agentic AI technology.

Agentic AI will create a new set of technical achievements and ethical dilemmas. While agentic AI can speed up processes and personalize customer experiences, these complex autonomous systems can produce unexpected consequences in the pursuit of organizational goals. Companies will need to get a handle on these increasingly powerful technologies.

Rosenfeld refers to Amara’s Law, which holds that societies tend to overestimate the effect of a technology in the short run while underestimating its impact in the long run.

“Right now, some people might believe AI is hyped or that the effect of it is not going to be as pronounced as some other technologies,” Rosenfeld says. “But over time, AI is going to surpass expectations. It will be even more important to businesses going forward.”

The middle market has adopted AI in the hopes of increasing productivity, efficiency and innovation. Organizations see the power of AI and are embracing it.

“Companies are adapting to AI,” Franko says. “And people are adapting too. Individuals can choose to be worried about AI taking their jobs, even if AI will likely create as many jobs as it eliminates. Or people can be proactive and take advantage of AI so that it doesn’t take advantage of them.”

Here, he discusses the role of AI in the middle market. 

Q: How has AI changed the way the middle market has operated over the past few years?

A: Artificial intelligence is becoming increasingly embedded in day-to-day operations across the middle market. It’s no longer just about automating tasks—though AI does reduce labor costs, increase operational efficiency and free teams from repetitive work. Today, AI is becoming a strategic asset, helping organizations make faster, data-driven decisions, streamline processes and gain a competitive advantage. As AI tools grow more sophisticated, they’re driving greater agility, enhanced scalability and deeper insights.

Q: What are the most significant AI-driven innovations for businesses?

A: Some of the most meaningful innovations stem from AI’s ability to learn from data and act in context. Tools that combine machine learning, natural language processing and automation are reshaping how work gets done. We’re also seeing early applications of agentic AI—such as systems that can initiate workflows or make routine decisions—allowing teams to focus on more strategic initiatives.

Q: Have there been any unexpected consequences of AI adoption for businesses?

A: Some businesses may be underestimating the level of change AI brings—not just in terms of upskilling staff, but in how teams interact with AI-driven tools. As these systems begin to operate more independently, organizations are finding that traditional workflows don’t always translate. Others are encountering issues with data quality, system integration or establishing proper governance to ensure outputs are accurate. While AI has improved speed and efficiency, it has also introduced new questions about accountability, especially since AI actions originate from systems rather than people.

Q: What are the biggest challenges or risks associated with AI adoption?

A: Many organizations are adopting AI, but unlocking real return on investment requires addressing key challenges—particularly around data quality and workforce readiness. Clean, standardized and complete data is essential for accurate, reliable models. AI is most effective when tailored to a company’s specific context and thoroughly tested to ensure performance and trust. There’s also a major opportunity to invest in change management and upskilling, helping employees embrace AI as a tool to enhance their work. With the right foundation and thoughtful integration, AI can drive smarter decisions, greater efficiency and a strong competitive edge.

Q: Has AI led to either job displacement or job creation in the middle market?

A: While AI may result in job displacement within certain functions, it is also leading to job transformation. Certain tasks, such as data entry or basic reporting, are increasingly being automated, and this has allowed human talent to focus on higher-value work, like strategy, creative ideas, client relationships and nuanced decision making. AI has also led to the creation of newly critical roles for data scientists, AI specialists and tech-savvy consultants, who can integrate AI tools into client solutions and help companies implement and adapt to new capabilities.

Q: Has AI influenced customer experiences or engagement? If so, how?

A: AI is significantly influencing customer engagement by helping organizations deliver faster, more efficient and personalized experiences. AI-powered tools analyze client data to generate tailored recommendations, streamline interactions and enable quicker turnaround times. This leads to more transparent communication and higher satisfaction. By automating routine tasks and enhancing service delivery, AI allows companies to focus more on strategic, value-added interactions—ultimately improving both the customer experience and long-term engagement.

Q: How are companies balancing AI implementation with maintaining human oversight?

A: Companies are leveraging AI for tasks that are repetitive or data-intensive, while keeping humans in the loop for complex or judgment-based activities. Regardless of what AI delivers, human professionals must still review the outputs and make nuanced decisions based on specific circumstances. People have to provide oversight and ensure that AI-driven insights align with regulatory requirements and professional standards, as well as the organization’s strategies, values and desired customer experience. It’s a collaborative model that helps mitigate risks, enhances performance and ensures that AI complements human judgment rather than replaces it.

Q: What advice would you give to professionals looking to future-proof their careers in an AI-driven landscape?

A: Professionals should focus on becoming adaptable and committing themselves to continual learning. First, they should invest in developing data literacy—understanding how to interpret and leverage data, because it will be at the heart of future AI tools. Professionals must also learn how to collaborate with these systems rather than compete with them. Gaining proficiency in AI-related tools, even at a basic level, will be increasingly important, along with understanding the ethical implications and governance frameworks around AI use. Finally, honing interpersonal skills—such as client communication, leadership and critical thinking—will remain invaluable, because AI cannot replace the human touch. 

Q: How do businesses handle ethical concerns, such as bias or data privacy, when using AI?

A: Companies should use diverse datasets and regularly audit AI systems to check for unintended outcomes. Organizations should also emphasize transparency and ensure that AI solutions align with ethical standards. Regarding data privacy, middle market businesses must implement robust data governance frameworks and strive to keep their customers’ personal information secure. Successful companies foster a culture of accountability and encourage teams to consider the ethical implications of AI, while engaging with external experts to ensure compliance and best practices.

Q: What opportunities do you think AI will create for businesses?

A: AI will create a wide range of opportunities for businesses—from accelerating long-term profitable growth to improving operational efficiency. It will enable companies to leverage data more intelligently, drive innovation, automate routine tasks and deliver more personalized and scalable experiences.

Additionally, AI can support smarter decision making, enhance risk management and enable new revenue streams through product and service innovation. As AI technologies evolve, businesses that adopt them responsibly will gain a competitive edge in agility, performance and customer engagement.

Q: How do you see AI evolving over the next few years?

A: AI is evolving beyond its current applications, moving into more agentic AI systems, which will not only analyze data but also make autonomous decisions within set guardrails. For the middle market, this could mean AI systems taking on more operational tasks, from customer service to supply chain management, without constant human intervention. We’ll also see AI become more accessible and scalable, with smaller businesses implementing sophisticated AI tools without the significant upfront investment that has traditionally been required. The key shift will be toward AI democratization, where businesses of all sizes leverage AI-powered insights and automation to drive growth, improve efficiency and make faster, more informed decisions. At the same time, AI will become more intuitive, with user interfaces designed for nontechnical teams, making it easier for employees to adopt and benefit from these technologies. Many of these employees will build custom apps for specific purposes, even if they are not developers, because AI can help with that process. This will create AI tools that can perform tasks more effectively than ever.

Q: Anything to add on the topic of AI and the middle market?

A: While large enterprises have dedicated AI teams, middle market companies are often more agile—allowing faster adoption. To fully leverage AI, they need a strong digital foundation. They need clean, accessible data and leadership that supports innovation and workforce reskilling. Success depends on balancing technology with human skills like critical thinking and decision making. AI offers major benefits—automating tasks, cutting costs, boosting efficiency and enhancing customer experience—but the challenge lies in knowing where to start and how to deploy it effectively.

Methodology

^{The RSM Middle Market AI Survey 2025: U.S. and Canada received 966 responses (762 U.S. respondents and 204 Canadian respondents). Most data points reflected a sample of 877 responses (691 U.S. respondents and 186 Canadian respondents). This represents the number of respondents whose organization currently uses generative AI in business practices.}

^{The margin of error for the total sample (n=966) is ±3.2 percentage points at a 95% confidence level. Due to rounding and multiselect questions, percentages may not total 100%.}

^{Unlike the 2024 study, for this year’s survey, respondents whose organization does not use AI tools (or who are unsure if they do) and also indicated their organization does not have an AI strategy and has no immediate plans to develop one (or are unsure about a formal strategy) were terminated from the study.}