Finance leaders have maintained a steady pace of AI adoption through 2025, even as early momentum has cooled and organisations grapple with data, skills, and governance challenges. Gartner’s latest AI in Finance Survey shows that 59% of CFOs and senior finance leaders now use AI within their finance functions, compared with 58% in 2024. The figure indicates consolidation rather than acceleration after last year’s sharp jump from 37% in 2023.

The poll, conducted in May and June 2025 among 183 senior finance leaders, suggests that finance teams have moved past the early wave of experimentation into a more measured phase, where practical obstacles and organisational readiness shape the speed of progress.

“The momentum in finance AI adoption has slowed following a sharp increase from 37% in 2023 to 58% adoption last year,” said Marco Steecker, Senior Director, Research in the Gartner Finance practice. “While overall growth in finance function adoption is proceeding more slowly, 67% of those using AI in finance are more optimistic about AI than they were last year. Notably, optimism tends to grow with AI maturity, with 23% of those farther along in AI adoption saying they were ‘much more optimistic’ about AI in 2025 compared to just 7% of those just starting to adopt AI.”

For corporate treasurers, this steadier pace reflects where many teams find themselves: aware of AI’s potential to support forecasting, anomaly detection, liquidity visibility, and process automation, but still limited by fragmented data structures and varying levels of digital maturity across the wider finance organisation.

Steecker notes that confidence tends to increase once organisations push through early constraints. “This growing confidence shows that even though the pace of AI adoption in finance has slowed due to complexity, data, and talent challenges, organisations that overcome these barriers are reaping significant rewards. As AI’s performance improves, and the technology evolves to address a wider array of use cases, CFOs should experience a virtuous cycle between AI development in finance and new opportunities to leverage the technology.”

That sentiment echoes a broader trend facing treasury teams: once core data foundations are established, the benefits of automation and predictive analytics compound quickly across cash, risk and working capital workflows.

Established AI use cases dominate early adoption

Gartner’s findings indicate that the finance sector is consolidating around a few well-understood AI applications that support efficiency and integrity in financial operations. Three use cases now stand out, with adoption by more than a third of organisations already using AI in their finance functions.

Knowledge management is the most common, used by 49% of respondents (see Figure 1). This includes AI tools that organise, retrieve and surface information to speed up decision-making, close knowledge gaps, and reduce dependency on manual document searches.

Source: Gartner (November 2025)

 

Accounts payable automation follows at 37%, reflecting the continuing push to streamline invoice processing, matching, and approvals. AI-driven anomaly detection stands at 34%, supporting early identification of irregularities across payments, reconciliations, and general ledger entries.

These use cases align closely with the workflows where corporate treasurers often see the first tangible gains. Automated data classification improves cash visibility. Anomaly detection reduces fraud exposure and strengthens payment controls. Natural language tools reduce cycle times associated with research and reconciliation.

Gartner notes that additional opportunities are emerging for teams that have reached a more advanced stage of adoption. “Some less-established, high-feasibility use cases also show significant potential. For example, the AI use case ranked highest for impact by finance leaders was code generation, by a significant margin,” said Steecker. “Organisations are finding this use case enables staff to find custom high-leverage opportunities for increased automation and insight generation.”

For treasurers, this could translate into dynamic forecasting models tailored to a specific organisation’s cash flow patterns, automated scenario generation, or custom data connectors that streamline integration between treasury systems, ERP modules and banking channels.

Barriers continue to slow large-scale adoption

The survey points to two factors behind the slowdown in overall adoption. First, a persistent minority remain unconvinced, with 16% reporting no planned AI implementations over the coming year. Second, a larger group, representing 25% of organisations, continue to struggle with the move from planning to piloting. This gap reflects uncertainty over use case prioritisation, data readiness, and the governance needed for controlled deployment.

Data literacy and technical skills remain the most widely cited barriers, followed closely by concerns over the quality and availability of data. These issues are especially relevant for treasury teams reliant on multiple banking partners, legacy connectivity and non-standardised cash flow data. Without a clean and consistent data foundation, AI models can deliver inconsistent outputs or require continuous rework.

Cultural acceptance also continues to hold some organisations back. Gartner’s findings show that finance teams with no AI adoption often struggle to establish organisational support or to frame AI as a strategic priority rather than a technology experiment.

Even for organisations that have launched pilots, significant impact takes time. According to the report, 91% of respondents initially experience low or moderate benefits, reinforcing the need for long-term investment and continued iteration. Results improve substantially as organisations progress. Those further along the adoption curve are more than twice as likely to report moderate impact and nearly three times more likely to see high impact, while being half as likely to report low impact.

The implications for treasurers are clear. Early pilots typically focus on narrow tasks, such as automated reconciliations or data extraction. As teams scale the technology and improve data quality, AI’s impact expands into forecasting, funding decisions, liquidity planning and risk analytics.

Scaling AI beyond early wins

Steecker argues that organisations should prioritise moving high-potential projects more quickly through development, given the compounding benefits seen at later stages of adoption. “Given that the biggest impacts are achieved once AI is in the production stages, finance leaders should focus on accelerating promising projects through early development,” he said. “AI’s potential in finance remains strong, and organisations that invest in overcoming adoption barriers will be well-positioned to capitalise on future opportunities.”

For treasurers, the findings reinforce a familiar message: successful AI adoption depends less on ambition and more on data readiness, skills, and the ability to translate insights into practical action. The steady adoption rate in 2025 suggests that finance teams are progressing cautiously but deliberately, setting the groundwork for more advanced applications over the next two to three years.

As AI tools become more embedded in core finance systems and cloud platforms, the gap between early adopters and those still evaluating their options may widen. The organisations that invest now in data quality, process integration and skills development will be better placed to capture the benefits once AI capabilities mature across forecasting, liquidity planning and risk management.