Each month, the Richmond Fed surveys businesses across its region to provide perspectives on current business activity and expected changes in business conditions.
Capital and AI Investment in the Fifth District
According to the Bureau of Economic Analysis, much of the recent growth in private sector investment in the U.S. economy has been technology related, from investment in information processing equipment and intellectual property products to exponential growth in data centers. In addition, many firms have been spending on services like subscriptions or training to determine how they can best employ artificial intelligence (AI) technology. In June, we asked over 250 firms in the Richmond Fed's Fifth District about their capital investment, AI adoption, and factors such as employment and prices. While there have been high levels of both capital investment and AI adoption in recent months, most firms are still determining how to use AI in their businesses and where or when gains from the technology will be realized.
AI Adoption Is Widespread
Both employees' access to AI and AI use in business operations were widespread in June. Among respondent firms, 89 percent indicated that their employees had access to AI tools for at least one of their regular tasks. Further, 83 percent of respondent firms reported that they used AI in one of their formal business operations.
AI is most commonly used in tasks such as summarizing, writing, and data analysis. The most frequently reported business applications are in sales and marketing, management, and strategy. Notably, operations which require the use of AI to interact directly with consumers — such as customer service — were less common than internal operations.
Firms Continue to Invest
As AI adoption and use increase, its impact on investment, prices, and employment becomes all the more important to understand. To better assess this dynamic among our surveyed firms, we examine reported growth in respondents' capital investment and the extent to which that investment is dedicated to AI.
All sectors reported positive average growth in their capital investment in the past 12 months. Firms in construction and professional and management services reported the largest average capital investment growth of 5.5 percent and 4.0 percent, respectively.
Few firms reported significant shares of their overall capital investments in AI. Among firms that had growth in their capital investments last year, 40 percent did not invest in AI, and a further 30 percent stated that AI investments made up only 1 percent to 9 percent of their total capital investments. In the question, we defined capital investments as "investments in equipment, software, technology, hardware, facilities, or other major purchases [including] implementation or maintenance of AI," but some firms may still categorize AI investment as operational or other spending. According to a computer services firm from South Carolina:
"Our AI investments consist of ... license costs, which is not a significant amount compared to overall capex spend, and also developing our own internal AI tools that avoid high token costs. Fortunately, we had the staff and network infrastructure to handle it. Our AI spend [in this question] doesn't reflect the significant investment in time we're making to develop those tools."
AI Spending as a Share of Investment: More on the Way
Although capital investments represent only a share of the total resources firms dedicate to their AI tools, firms across sectors plan for AI-related investment to account for a greater share of their capital investment next year relative to this year. Looking at expectations over the next 12 months by industry, the professional and management services sector expects the largest share of investment to be AI-related; Twenty-four percent of firms indicated that they will invest 10 percent to 24 percent of their capital investment into AI; 14 percent indicated that they will invest 25 percent to 49 percent; and 7 percent indicated they will invest 50 percent to 74 percent of their capital investment. Other surveys, such as the Business Trends and Outlook Survey and The CFO Survey, also indicate that the high-skilled service sector is both experiencing and expecting the largest investment.
Insights From Firms
Some Fifth District businesses have already reported efficiency improvements due to AI. Most, however, are still exploring how AI will fit into their business processes. For example, some respondents are in the planning phase, some are developing their tools, and some are just now allowing their employees to experiment with AI tools. One wood and pallet manufacturing firm from North Carolina stated:
"We are not yet able to implement AI in our production process, although we are actively trying to implement it in our planning and scheduling processes for production. Currently, a handful of us use LLMs to help with idea creation, docs, etc."
In addition, a commercial real estate firm from North Carolina commented:
"We are just starting to allow employees to download Claude and Copilot to do analytical work for project analysis in real estate. We use it for lease abstract and proposal responses. We are just starting to train on how to ask AI questions for the product we want it to produce."
Respondent firms indicated that efficiency effects from AI may be imminent. One firm in the technology sector noted:
"[AI] makes the cost of developing software very low. It makes marketing easier and cheaper."
Moreover, an internet marketing services firm in Virginia stated:
"[Our] shift to AI in workflows is dramatically increasing productivity, though not without a lot of internal investment in governance, training, tooling."
Some firms also reported the impact of AI on employment — a question on the minds of researchers, policymakers, employers and workers. Very few firms said they expect a net negative impact from AI on employment, and one Virginia technology firm reported a positive impact from AI on employment. This technology firm from Virginia stated:
"We use AI to a great extent in the development of software ... and in performing our day-to-day business operations, and while our use of AI continues to increase, we have not decreased our staff. On the contrary, we find that the more we are using AI, the more productive our employees can be, and the more output we can produce, but that we still need the same number, if not more humans in the loop to perform different tasks. For example, we can generate more software code in shorter amounts of time, but then we need more humans to create the prompts that instruct AI to create the code, more humans to check and, test that code, and more humans to interact with our customers to ensure the code is meeting their needs."
According to this firm, AI-driven productivity increases require more, not fewer, employees. The impact of AI on firms' employment expectations is a topic we will continue to explore through our business surveys.
Conclusion
Both quantitative and qualitative evidence from Fifth District firms indicates there has been a rise in AI spending, but the use of AI tools is still young. Although individual firms commented that AI increased efficiency, we have not found evidence across firms that AI has increased productivity at scale. It might take some time for the impacts of AI on Fifth District firms, and the regional economy, to become clear.
Views expressed are those of the author(s) and do not necessarily reflect those of the Federal Reserve Bank of Richmond or the Federal Reserve System.
Your participation in our important monthly business survey ensures Federal Reserve leadership have accurate and timely economic data.
