September 2025
PYMNTS Data Books

AI Agents in the Enterprise

CFOs are building and buying, but every fully automated action right now demands human approval.

Get Unlimited Access
Complete the form below for free, unlimited access to all our Data Studies, Trackers, and PYMNTS Intelligence reports.

Thank you for registering. Please confirm your email to view all our Trackers.

    yesSubscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    Agentic AI is starting to move from concept to deployment in corporate finance. In a PYMNTS Intelligence phone survey of 60 verified chief financial officers at U.S. companies with at least $1 billion in revenue last year, roughly one in 10 report using or piloting AI agents. Engagement reaches 35% when you include firms exploring or considering the software.

    Fielded from July 15, 2025, to July 25, 2025, the survey shows varied approaches. Of those already using agentic AI or piloting/testing the software, 71% are building technologies in-house. Among those exploring adoption, 43% said they would build in-house. Technology firms, the inventors of agentic AI, lead the adoption curve. CFOs use the software for financial planning and analysis (FP&A) and financial reporting functions first, while treasury and risk processes rank lower. Among current adopters, 43% would grant agents “moderate” system access with human approval. As of July 2025, none would allow substantial or full autonomy.

    The current state of agentic AI adoption

    One in eight firms (12%) is already using or piloting agentic AI. Twenty-three percent are engaged in exploring the software, and 65% have no plans.

    In July 2025, 11.7% of large enterprise CFOs said their companies were using or piloting agentic AI. Tech firms lead, with 40% using/piloting, versus 12% in services and 13% in goods.

    Gen AI use impacts agentic AI use

    Half of “high impact” gen AI companies, meaning those with significant gen AI applications carrying significant business value and risk, are using or pursuing agents. For “low impact” companies, 83% have no plans.

    Enterprises that already see gen AI as high impact are twice as likely to be moving to agentic AI. One-quarter of this cohort are already using or piloting the technology. At the same time, most low-impact peers remain on the sidelines, with 82.6% reporting no plans.

    Approaches to using agentic AI

    Build wins: 71% of active adopters of agentic AI are building in-house; 43% buy off the shelf.

    Among companies already using or piloting agents, the dominant approach is standing up internal AI/engineering teams. Nearly half of the firms surveyed also procure packaged tools, signaling a hybrid “build and buy” posture as projects scale.

    Usage of agentic AI for specific tasks

    The majority of firms are very interested in FP&A use cases; only 32% say the same for treasury functions.

    CFOs put planning and reporting first. Seven in 10 are very or extremely interested in FP&A, 68% in financial reporting and 63% in cost, working capital and strategy support. Interest drops for treasury (31.7% are very/extremely interested). Risk (45%) and governance/compliance functions (46.7%) sit in the middle of the pack.

    Level of access

    Access is the gate: 43% of active adopters would grant “moderate” system access; none would allow full autonomy. Agent rollout hinges on permissions. Among firms already using or piloting agentic AI, 43% would let agents tap internal directories and execute tasks with human approval. But as of July 2025, no CFOs would grant substantial or full access.

    Conclusion

    Adoption of agentic AI is early and cautious, and it is concentrated in firms that already see gen AI as high impact on their operations. Half of the high-impact cohort is using, piloting, exploring or considering agents, while 83% of low-impact peers have no plans. Near-term deployment will sit in planning, reporting and cost work, with hybrid build-and-buy strategies and human-in-the-loop controls. Wider rollout hinges on permissions and governance. Today, no CFOs say they are ready to grant substantial or full autonomy.

    About

    PYMNTS INTELLIGENCE

    PYMNTS Intelligence is a leading global data and analytics platform that uses proprietary data and methods to provide actionable insights on what’s now and what’s next in payments, commerce and the digital economy. Its team of data scientists includes leading economists, econometricians, survey experts, financial analysts and marketing scientists with deep experience in the application of data to the issues that define the future of the digital transformation of the global economy. This multi-lingual team has conducted original data collection and analysis in more than three dozen global markets for some of the world’s leading publicly traded and privately held firms.

    We are interested in your feedback on this report. If you have questions or comments, or if you would like to subscribe to this report, please email us at feedback@pymnts.com.

    Disclaimer

    PYMNTS Data Books may be updated periodically. While reasonable efforts are made to keep the content accurate and up to date, PYMNTS MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED, REGARDING THE CORRECTNESS, ACCURACY, COMPLETENESS, ADEQUACY, OR RELIABILITY OF OR THE USE OF OR RESULTS THAT MAY BE GENERATED FROM THE USE OF THE INFORMATION OR THAT THE CONTENT WILL SATISFY YOUR REQUIREMENTS OR EXPECTATIONS. THE CONTENT IS PROVIDED “AS IS” AND ON AN “AS AVAILABLE” BASIS. YOU EXPRESSLY AGREE THAT YOUR USE OF THE CONTENT IS AT YOUR SOLE RISK. PYMNTS SHALL HAVE NO LIABILITY FOR ANY INTERRUPTIONS IN THE CONTENT THAT IS PROVIDED AND DISCLAIMS ALL WARRANTIES WITH REGARD TO THE CONTENT, INCLUDING THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT AND TITLE. SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OF CERTAIN WARRANTIES, AND, IN SUCH CASES, THE STATED EXCLUSIONS DO NOT APPLY. PYMNTS RESERVES THE RIGHT AND SHOULD NOT BE LIABLE SHOULD IT EXERCISE ITS RIGHT TO MODIFY, INTERRUPT, OR DISCONTINUE THE AVAILABILITY OF THE CONTENT OR ANY COMPONENT OF IT WITH OR WITHOUT NOTICE.
    PYMNTS SHALL NOT BE LIABLE FOR ANY DAMAGES WHATSOEVER, AND, IN PARTICULAR, SHALL NOT BE LIABLE FOR ANY SPECIAL, INDIRECT, CONSEQUENTIAL, OR INCIDENTAL DAMAGES, OR DAMAGES FOR LOST PROFITS, LOSS OF REVENUE, OR LOSS OF USE, ARISING OUT OF OR RELATED TO THE CONTENT, WHETHER SUCH DAMAGES ARISE IN CONTRACT, NEGLIGENCE, TORT, UNDER STATUTE, IN EQUITY, AT LAW, OR OTHERWISE, EVEN IF PYMNTS HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
    SOME JURISDICTIONS DO NOT ALLOW FOR THE LIMITATION OR EXCLUSION OF LIABILITY FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES, AND IN SUCH CASES, SOME OF THE ABOVE LIMITATIONS DO NOT APPLY. THE ABOVE DISCLAIMERS AND LIMITATIONS ARE PROVIDED BY PYMNTS AND ITS PARENTS, AFFILIATED AND RELATED COMPANIES, CONTRACTORS, AND SPONSORS, AND EACH OF ITS RESPECTIVE DIRECTORS, OFFICERS, MEMBERS, EMPLOYEES, AGENTS, CONTENT COMPONENT PROVIDERS, LICENSORS, AND ADVISERS.
    Components of the content original to and the compilation produced by PYMNTS is the property of PYMNTS and cannot be reproduced without its prior written permission.