Gen AI accelerates transformation in finance as pilot success remains elusive

Despite rapid growth in generative AI adoption among CFOs, few pilots deliver measurable profit impact, prompting a shift towards embedding AI in core finance functions for strategic advantage.

Artificial intelligence (AI), particularly generative AI (gen AI), has become a transformative force in finance functions, reshaping how businesses manage processes, insights, and decision-making. According to an extensive McKinsey survey of 102 CFOs globally, adoption of gen AI ...

for multiple finance use cases surged from 7 percent to 44 percent between 2024 and 2025. Additionally, 65 percent of organizations expect to increase investment in gen AI this year, marking significant growth compared to just a quarter two years ago. Despite these promising trends, however, only a small fraction of AI pilots—about 5 percent—have achieved meaningful profit-and-loss impact. This gap often stems from the difficulty of scaling pilots beyond controlled environments and integrating AI into core workflows effectively.

Successful finance teams are those moving beyond isolated experimentation toward embedding AI in foundational finance functions, including strategic planning, cash and working capital management, and cost optimisation. AI-powered decision support tools, employing predictive analytics combined with gen AI capabilities, are enabling faster, deeper insights. For instance, a global consumer goods company uses a gen AI assistant to analyse budget variances across divisions, freeing finance professionals from manual data crunching and saving around 30 percent of their time. Similarly, a global biopharma firm employs a gen AI and agentic AI-assisted decision support agent to halve the time needed for resource allocation decisions by generating complex scenarios in natural language, integrating multiple data sources such as financials and marketing analytics. These AI tools not only alert finance teams to performance issues—such as overspending or declining ROI in specific channels—but also provide root-cause analyses and data-driven action recommendations.

Agentic AI, an emergent category characterised by autonomous goal pursuit and decision-making, is gaining traction in automating workflows like accounting close processes or complex report drafting. A major North American financial institution leverages gen AI to draft risk model reports and develop market-specific risk models by combining internal and external data, drastically reducing time and effort.

AI is also revolutionising cash and working capital management by scrutinising contracts and invoices with greater accuracy. For example, a global biotech firm implemented an agentic AI system to monitor invoice-to-contract compliance across the year. This tool detects errors such as missed early payment discounts or volume rebates by interpreting contract terms and tracking invoice alignment, uncovering contract leakage amounting to approximately 4 percent of total spend—a margin recovery opportunity worth tens of millions for large-scale companies.

Cost optimisation benefits from AI’s ability to categorise sprawling and complex spend data. A large European financial institution employed large language models and advanced analytics to organise supplier invoices into a detailed taxonomy with hundreds of subcategories, revealing cost waste in areas like energy usage, travel, and facilities. This identification of inefficiencies led to a cumulative cost reduction of about 10 percent across a multibillion-euro spend base. Another European packaging company used gen AI to classify over 10,000 suppliers, uncovering overlooked opportunities for cost savings and supplier diversity expansion.

However, realising AI’s full potential requires finance teams to overcome several hurdles. Common pitfalls include waiting for perfect data before implementation, attempting wholesale transformation rather than incremental domain-by-domain change, launching pilots without clear strategic roadmaps, neglecting change management, and automating fragmented processes without first simplifying workflows. Delays in adoption often arise not from technological limitations but from organisational resistance and lack of alignment. Industry experts advocate for piloting AI solutions with existing imperfect data, building momentum through targeted use cases that address business priorities, and fostering adoption through training and transparent communication.

Beyond these practical challenges, integration with legacy systems and ensuring data security and privacy remain critical. Reports from CFO-focused sources and AI in finance specialists highlight that legacy infrastructure underscores complexity in deploying AI tools, often necessitating technical upgrades along with process reengineering. Robust encryption, access controls, and compliance with data privacy regulations are vital to maintaining trust and safeguarding sensitive financial information.

Academic advances further enrich this evolving landscape. Recent research papers introduce sophisticated AI-native, agent-based frameworks designed for enterprise resource planning (ERP) in finance, such as FinRobot. These architectures blend generative AI with business process modelling and multi-agent orchestration, enabling end-to-end automation of complex tasks—from budget planning to wire transfers—demonstrating substantial reductions in processing time, error rates, and improved regulatory compliance. Open-source platforms focused on AI agents powered by large language models (LLMs) aim to democratise access to advanced AI capabilities for financial decision-making, contextualising AI as a tool not only for automation but for enhancing strategic analysis.

The evolving consensus among finance leaders and AI researchers is clear: the opportunity AI represents is real and growing, but capturing sustained value demands moving beyond pilots to disciplined, strategy-driven execution. Finance functions that integrate AI into core processes, invest in data and technology foundations, and cultivate an adaptive culture stand poised to become more agile, insightful, and aligned with broader organisational goals. In this way, AI is transitioning from a futuristic promise to an essential element of effective financial management and strategy in the digital age.

Source: Noah Wire Services