AI has dominated business headlines for the past two years, and finance functions are no exception. In a proprietary McKinsey survey of 102 CFOs across industries and global regions, 44 percent of respondents said they used gen AI for over five use cases in 2025, up from 7 percent in the previous year’s survey. Investment in AI tools is also growing: 65 percent of respondents said their organizations will increase gen AI investment in 2025; two years ago, only about a quarter of respondents said the same.
Yet the reality across the corporate landscape underscores how elusive tangible value remains: According to one study, only about 5 percent of AI pilots have translated into meaningful P&L impact.1 Poor outcomes are largely due to pilots breaking down under real-world conditions, failing to adapt as new data emerges, and remaining poorly integrated into core processes.
Some finance teams, however, are successfully using AI, gen AI—and increasingly, agentic AI—to boost efficiency, improve insights, and offload time-consuming manual tasks (see sidebar, “A guide to automation and AI terms.”) Rather than relying on isolated pilots, these organizations apply AI across foundational finance domains. We have observed some CFOs and their teams using AI to forecast more accurately, monitor working capital in real time, speed up reporting cycles, and surface new opportunities for cost savings. These efforts are enabling them to become more agile, forward looking, and aligned with the needs of their organizations.
This article examines three areas where, based on our experience, finance teams are delivering the most value with AI: strategic planning and control, cash and working-capital management, and cost optimization. Each section includes case studies that illustrate how leading organizations use gen AI and agentic systems to improve how finance functions operate. Finally, we identify five common missteps that can slow things down and what it takes to get past them.
Strategic planning and control: How AI can deliver better insights
Decision support tools, powered by a combination of predictive analytics and gen AI, make it faster and easier to access company data, generate reports, and run forecasts or scenarios. These tools support finance leaders and their teams while also making data more accessible to decision-makers across the business. Typically, AI tools combine a few common capabilities: alerts that help finance leaders focus their time and attention, interactive root-cause analysis that helps the user understand what’s influencing performance, and alternative scenarios to consider. AI is suited for these tasks because it’s particularly good at integrating multiple layers of information—such as from external, financial, and operational sources—into a cohesive view.
For example, at a global consumer goods company, a gen AI assistant helps finance professionals deliver insights on budget variances to business leaders in different divisions and markets. The tool replaces manual number crunching, saving an estimated 30 percent of finance professionals’ time.
In another example, a global biopharma company’s decision support agent, enabled by gen AI and agentic AI, cuts in half the time the finance team needs to make resource allocation decisions. Instead of manually pulling reports and stitching together insights across functions, the team now generates complex scenarios using natural language during monthly planning sessions. The AI tool integrates data from multiple sources—including customer-relationship-management systems, financials, and marketing mix analytics—to surface performance alerts (for example, when there is overspending against budget or declining ROI in specific channels). It also provides root-cause analysis (for example, “The problem stems from cost category A in region Y”). The tool then suggests data-driven action steps (for example, “Based on recent ROIs and forecasts, consider shifting 10 percent of the sales force budget to digital marketing to encourage higher growth”).
Finally, at a large North American financial institution, a gen AI tool helps generate first drafts of reports that document internal risk model requirements and updates. The tool also assists in generating market-specific risk models by combining internal data with public sources, streamlining what was once a time-intensive process.
Specific AI implementation varies by organization, of course. Across a handful of finance functions where it has been adopted robustly, we have observed that finance professionals spend 20 to 30 percent less time crunching data. They devote the saved time to their role as business partners who support strategy execution. By quickly generating customized reports that maintain appropriate security and hierarchical-access controls, AI tools also make it easier for finance departments to provide insights across their organizations.
Cash and working capital management: How AI scrutinizes terms and invoices for greater accuracy
AI-powered agentic workflows are enabling the next level of automation in both payable and receivable processes, helping make procurement and other back-office teams more efficient.
For example, a global biotech company introduced invoice-to-contract compliance using an agentic AI system that ingests contracts and invoices throughout the year and checks that all contract terms are correctly applied. This approach helps prevent value leakage when vendors miss or misapply terms such as early payment discounts, tiered pricing, and volume rebates. It runs alongside preexisting automation, extending coverage across the full range of the company’s spend base and reducing the need for manual monitoring of high-value contracts. The system is able to interpret each vendor contract and its terms, track incoming invoices for compliance, and identify issues that emerge only across multiple invoices, such as when cumulative purchase volumes trigger eligibility for a lower-priced tier.
By using this AI system, the company identified contract leakage equal to approximately 4 percent of total spend (a level of leakage that is not uncommon in the industry). This translated into a clear opportunity to recover lost value and improve margin performance. To put this in a hypothetical context, for a company with a nominal spend of $1 billion, closing that gap could result in a recurring margin improvement of $40 million.
Cost optimization: How AI finds savings by analyzing granular spending
AI can simplify the time-consuming task of categorizing detailed costs by analyzing complex invoices and purchase orders and organizing them into clear, structured categories. With better visibility, finance teams can apply advanced algorithms to spot anomalies and areas of waste.
To better understand and control its indirect-spend base, a large European financial institution set out to identify hidden inefficiencies across its operations. It began by collecting invoice-level data from thousands of suppliers and organizing it into a detailed cost taxonomy with four levels of increasing detail and approximately 400 subcategories. To efficiently process and classify this data, the organization used a combination of large language models and advanced analytics. With the structured data set in place, it surfaced cost inefficiencies by applying both automated and semiautomated methods (for which an expert reviewed the output) for anomaly and pattern detection. This analysis revealed specific opportunities to reduce costs and cost waste in areas such as energy usage, travel and transport, and facility management. While each category delivered modest savings on its own, together they helped reduce costs by approximately 10 percent of a multibillion-euro spend base.
Another large European company in the packaging industry gained better control over a fragmented supplier base by using gen AI to categorize more than 10,000 suppliers. Management had historically focused on top-spend vendors, while numerous smaller suppliers—many in indirect-spend categories—remained poorly understood. Using gen AI, the company classified all suppliers with greater accuracy, identifying patterns and overlaps that had previously gone unnoticed. This enhanced visibility helped uncover cost-saving opportunities and optimize procurement strategies. The categorization also revealed gaps in supplier diversity, enabling the company to expand sourcing in underserved areas.
Overcoming barriers to scaling AI in finance
To capture AI’s potential in finance, teams will need to do more than add new tools on top of old ways of working. They must rewire core processes, talent, and technology so that adoption takes hold and creates value. Along the way, progress can be slowed or stalled by these common pitfalls:
- Waiting for perfect data. Some teams delay rewiring processes until every data set is perfectly accurate, connected, and standardized. In practice, finance teams can create value by delivering use cases that work with today’s data while also strengthening data foundations.
- Trying to transform all at once. Holding back until the entire function is “AI ready” slows progress. The better path is to transform domain by domain, building momentum and capabilities that deliver sustainable results.
- Jumping in without a clear road map. Pilots launched without direction rarely scale. Finance leaders need a road map tied to their business priorities, with clear choices of which use cases to pursue first and which to take on next. Use cases should also be supported by the technical talent that can help them succeed.
- Neglecting change management. The biggest barrier is often adoption, not technology. Equipping teams and building buy-in are essential for capturing and sustaining impact.
- Automating fragmented processes. Without simplifying and standardizing core workflows first, AI only adds to the complexity. Removing unnecessary steps and making processes consistent across teams allows technology to scale effectively.
Avoiding these pitfalls requires a clear vision, strong business alignment, and a focus on practical execution. Finance leaders who approach AI with a strategy rooted in business needs are best positioned to achieve enduring impact.
As AI adoption broadens, the difference between pilots that fizzle and those that create lasting value is becoming clear. As the case studies in this article show, the companies achieving results are the ones that tie AI to specific business needs, streamline core processes, and use the technology to free up capacity for higher-value work. For CFOs, the message is unequivocal: The opportunity is real, but capturing it requires moving beyond experimentation to disciplined execution anchored in business priorities.
