Finance teams are investing heavily in artificial intelligence, but the results are showing up a bit scattered and in different places.
A global study from KPMG of more than 1,000 senior finance leaders finds AI is already embedded across financial planning, reporting and analysis. A separate survey of 311 U.S. accountants, commissioned by Progress ShareFile, a provider of document management and workflow software, shows similar adoption among accountants with AI now embedded in day-to-day accounting work.
Some are seeing gains in decision-making, forecasting and responsiveness, while others are focused on improving workflows and reducing the number of steps it takes to get work done. Across both survey sets, adoption is widespread, but the outcomes vary based on how the technology is being applied.
The results also point to a major disconnect between where finance leaders are seeing gains and where accounting teams are still experiencing friction, and it’s all being shaped by how AI is applied across the function.
In the KPMG data, ROI is concentrated among organizations using AI within planning and forecasting to support decision-making. In the Progress survey, accountants point to workflow inefficiencies, process design and execution as ongoing challenges, suggesting that gains at the top of the function are not always translating into execution as the technology makes its way down the totem pole.
Adoption is broad, outcomes are uneven
AI use across finance and accounting has expanded quickly, and both datasets show it is now embedded in many companies’ core workflows. KPMG data shows more than three-quarters of organizations are using AI across FP&A, while nearly three-quarters (71%) say it is meeting or exceeding ROI expectations across the finance function. Progress data shows similar adoption at the practitioner level, with more than four in five (84%) reporting AI use in their work.
Even with that level of adoption, positive ROI is concentrated among a smaller group of organizations. In the KPMG data, fewer than one-quarter (23%) say AI is exceeding expectations.
At the same time, many accounting teams describe day-to-day work shaped by inefficiencies tied to fragmented workflows, disconnected tools and a lack of automation in routine processes, conditions that align with the data quality, governance and integration gaps KPMG identifies as central to getting real performance gains from AI.
In the Progress data, more than half (57%) of accountants say they can perform their jobs effectively but not efficiently, pointing to some core processes that they say slow them down. That friction shows up in how work moves across systems, as nearly two-thirds (65%) cite a lack of automation in routine procedures, while more than six in ten (61%) say switching between tools remains a barrier, suggesting that tasks are still fragmented rather than moving through a unified process.
Three-quarters (75%) also report workflows that include too many steps, reinforcing how incremental tasks and handoffs continue to determine execution across accounting teams, challenges KPMG links to broader issues around data governance and integration during AI adoption.
That distribution also reflects how results are taking action across the function. In the KPMG data, strong performance is concentrated among organizations applying AI within finance processes tied to planning and forecasting. In the Progress data, accountants continue to point to workflow design, system integration and process flow as areas where work slows down. The pattern shows how adoption is moving across the function, while outcomes are influenced by how AI is deployed and how it interacts with existing systems.
The variation in outcomes also strongly aligns with how finance teams are approaching implementation. In the KPMG data, organizations that have embedded governance, measurement and oversight into their AI deployments report stronger performance improvements.
This dynamic is interesting when taking into consideration that accounting firms, somewhat ironically, have presented themselves as technology-forward while still working through inefficiencies in practice, a gap that continues to surface as AI adoption expands.
Where value is showing up
The variation in outcomes is also reflected in where AI is having the most measurable impact across the finance function. In the KPMG data, 70% of organizations report improvements in decision quality, while a similar share (71%) cite gains in decision speed and nearly two-thirds (64%) point to improved forecast accuracy. These gains are concentrated in areas where finance teams are working with structured data and defined processes.
The concentration of value in these areas is tied to how AI is being applied within core finance activities. Organizations that are further along in their AI deployments are using it to support decision-making cycles, where the ability to process large volumes of data and update models in real time directly affects outcomes. These use cases tend to rely on consistent inputs and more clearly defined workflows, allowing AI to be applied more effectively.
At the same time, Progress data shows that many accounting teams are still working through how automation fits into execution. Nearly all respondents (98%) say they need better workflow technology, pointing to major gaps in how processes are structured and supported by leadership. That need is reflected in how work is carried out across systems, where bureaucratic layers like multiple steps, handoffs and approvals continue to shape day-to-day operations.
Slow operations, according to the Progress data, can seriously hinder a company’s ability to integrate these technologies at scale. More than one-third (34%) cite integration challenges as a barrier to automation, while a similar share (33%) points to security concerns, highlighting how system connectivity and data protection influence how AI can be deployed. Leadership resistance also plays a role, with 20% identifying it as a barrier, suggesting that adoption is not only a technical issue but also an organizational one.
These dynamics determine where value is emerging across the function. In finance, gains are emerging in areas tied to planning and analysis, where AI can be applied to structured data and repeatable processes. In accounting workflows, the focus remains on execution, with factors such as integration, security and process design shaping how work is performed. As a result, the impact of AI reflects how it is embedded within different parts of the finance function.
What comes next
The gap between where AI is being adopted and where it is delivering consistent results is increasingly tied to how finance functions are structured to support it.
In the KPMG data, organizations that report stronger outcomes are also those that have built the underlying conditions needed to scale AI, including clearer governance frameworks, more reliable data and defined processes around how decisions are made and executed. Those elements are determining factors in how AI is applied, measured and trusted across the function.
In accounting workflows, many of those same conditions are still being built. Progress data shows ongoing challenges tied to system integration, process design and how work moves across tools, all of which influence how effectively automation can be deployed.
As AI continues to become more embedded in the CFO’s daily workflow, those operational details take on a larger role in determining outcomes and what the rollout and eventual day-to-day role of AI in the finance function will look like.
