Finance functions are at an inflection point: many finance leaders have reached the limits of current capacity and performance while organizational and regulatory expectations for strategic contribution continue to grow. Over the past 15 years most finance teams have pursued a "do more with less" agenda — freeing time for strategic analysis, reducing effort on automatable tasks, and cutting costs by up to 25% of revenue share.
What worked — and what now limits progress
Organizations have pursued several strategies: adopting new technologies to reduce transaction and processing costs; creating global and on-shore business services models combined with technology; and linking different technologies with edge devices to make decision-making data-driven. The outcome has been cloud ERP systems, integrated data platforms, automation and edge analytics — in short, digital finance. Yet the persistent challenge is keeping pace with organizational and investor demands.
The breakthrough is a new way of working with AI agents
The next shift is not another system but a change in how work gets done. AI agents are reshaping work across industries; finance can — and in many cases already does — lead this change. AI agents operate intelligently and autonomously, can work in teams, and take on distinct roles such as accounting, FP&A, or compliance. According to the PwC AI Agent Survey, 79% of leaders reported that AI agents have been introduced at their organizations, and 34% use them in accounting and finance. Early evidence shows AI agents can deliver 60–90% short-term time savings on certain core processes.
How this works in practice
AI agents are provisioned with role-specific skills and appropriate datasets, then orchestrated into workflows. In invoice processing, for example, one agent extracts invoice key data, another retrieves the relevant contract, a third reconciles the invoice against purchase orders or contract terms and flags discrepancies, while a fourth drafts emails to resolve questions — only then does a human reviewer approve or escalate. In such cases agents can reduce cycle times by up to 80% while improving control trails and lowering compliance risk.
In treasury operations agents can consolidate cash balances, forecast short-term inflows and outflows, flag potential surpluses or shortfalls, recommend transfers or investments, log executed steps and refine forecasting models. Treasury teams then use those forecasts to refine capital allocation, adjust cash thresholds and update investment policies.
Other applications and scalability
Beyond accounts payable and treasury, agents can support collections management, supplier risk monitoring, liquidity optimization, financial closings and many other complex processes. They are readily scalable and finance team members can create or modify them; new agent workflows can reuse prior code and architectures, increasing speed to value.
Steps to build AI-enabled finance
- Assess your processes and data platforms.
- Build modularly and for reusability: modular architectures make it easier to reuse code, agents and agent frameworks across workflows, enabling rapid scaling and low costs.
- Recognize that a single agent typically performs one simple task; real value comes from connecting agents into workflows, which requires an easy-to-use platform such as the PwC agent OS with built-in governance tools.
- Demonstrate the employee benefits: the 2025 Global AI Workforce Barometer shows AI makes workers more valuable when they acquire new skills such as AI oversight and when teams adopt an AI-collaboration culture.
Risks and required enablers
The opportunities come with the need for updated governance frameworks, refreshed competencies and trust in operating models and AI solutions. Deploying AI agents is not only a technology initiative — it requires organizational change management, regulatory compliance and ongoing oversight.
Conclusion
Advances in artificial intelligence give finance functions a window to create capacity, automate routine tasks and free people for higher-value strategic work. To capture that potential, organizations need modular architectures, robust governance and targeted reskilling; without these, finance risks falling behind on delivering the strategic value stakeholders now expect. Our series will continue with deeper coverage of governance models, competencies and deployment patterns.
