{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.text}}
Key takeaways
AI investment has a measurable tipping point
Companies investing above 1.6% of revenue in AI reported stronger enterprise outcomes, including higher EBITDA growth, shareholder return, and revenue performance relative to peers.
Enterprise value comes from transformation—not technology spend alone
Organizations seeing stronger outcomes are redesigning workflows, modernizing operations, and embedding AI into core business processes instead of layering AI onto legacy ways of working.
Benchmarks help leaders move from experimentation to enterprise action
Financial, operational, workforce, and trust metrics can help executives assess AI performance, identify gaps, and make more informed decisions about where to scale, rebalance, or course correct.
If you’re in the C-suite or leading a line of business at your company, how can you know that you’re taking the right approach to AI, neither misallocating capital, settling for subpar returns, or falling behind? The answer is data. Benchmarks can help you answer questions like these:
Where, how, and how much should we invest in AI?
How can we get beyond pilots and deliver enterprise outcomes?
Are our AI investments delivering suitable financial and operational returns?
Are we creating enough stakeholder trust to enable AI value at scale?
Where should we rebalance, double down, or course correct?
Benchmarks by themselves shouldn’t define your strategy. But, by providing measures of your competitiveness across key dimensions, they can help you make better informed strategic decisions: where you should lead, where you can lag, and where you should exit. They can spot underperformance, risks, and opportunities, so leadership can pass the AI reality check and select where to focus AI efforts. Because advanced financial, operational, functional, trust, and workforce AI benchmarks don’t just look back. They can help you select, monitor, and evolve your path forward.
Based on PwC's 2025 AI Metric Survey of 70 senior business and technology leaders conducted from December 2025 to January 2026, and the analysis of publicly available financial information, here are the top takeaways (see methodology section below for additional details).
Value creation surges: The AI tipping point
If you double AI investments, you probably won’t get twice as much value. Depending on your starting point, you might get far less—or far more. According to our data, until companies cross an investment threshold, ROI tends to be insignificant. Above it, key enterprise outcomes can surge, according to our survey and analysis: EBIDTA up by 9.5%, total shareholder return up by 20.2%, and revenue up by 3.5%. (To avoid potential distortion, these numbers don’t include the AI foundation builders, where the AI bonus is far higher.)
This tipping point seems to be 1.6% of revenue, which puts companies in top quartile of AI investment. We call these firms AI leaders. Yet, this number won’t stay constant. An analysis of ours from last year, for example, found a lower tipping point. It’s also just an average. The right target for your organization may be different. But the main takeaway remains: Just throwing money at AI isn’t enough for business success. But if you invest too little, even the wisest moves can’t bring you the scale and capabilities needed for enterprise outcomes.
Investment patterns: balance AI business transformation and technology spend
If you layer AI onto old processes, you’ll probably just get slight improvements. Real value usually requires reinventing processes from end to end. You can then replace multiple, high-touch workflows with a single one that has AI at the core, doing most of the routine or repetitive work. People handle exceptions, make high-risk or high-value decisions, or shift to more strategic, creative, or interpersonal roles.
That’s likely why many companies are settling on a balanced investment mix: Direct tech spending is about three fifths (62%) of AI spend. Process redesign and change management is 34%, and training is 4%. These numbers are a cross-industry average. And as companies build their AI architecture and get through a big chunk of necessary tech spending, the “ideal” mix will evolve. Based on what we’re seeing in the field, for most companies, workforce transformation will be especially important: rethinking not just skills, but also roles, job architectures, incentives, pipelines, and recruiting. Either way, it’s crucial to keep track of how you’re allocating AI investments, so you’re not overemphasizing tech and neglecting transformation.
How to deliver enterprise-scale value: transform core operations
AI leaders put AI at the core of processes that matter. They report automating as much as half of customer interactions. Top performers have compressed cycle times in key finance processes (FP&A, procure to pay, order to cash, and record to report) by nearly two fifths. Some have cut IT incident resolution time by nearly three fifths.
Automated customer interactions and nimbler finance and IT functions are the kind of outcomes that keep costs down and enable business agility. And since they’re usually straightforward to measure, they make excellent benchmarks, so you can assess how well your AI is moving past pilots and into core operations.
How to boost adoption, performance, and sustainable ROI: trust metrics
How can you know how reliable and risk managed your AI is? Where do you need more human oversight? How good is good enough? Benchmarks can help answer these questions, so you can boost performance and the stakeholder trust that drives adoption and sustainable ROI. It’s critical that these benchmarks be grounded in real-world data, as our survey found a gap between realized performance inside the enterprise vs. what shows up on public leaderboards. This shouldn’t be surprising as the complexity of the modern enterprise can be hard to approximate with test data sets. Based on our survey data, here are four benchmarks that we find especially useful:
- Accuracy. How well do AI-generated insights align with verified data or expected outcomes? 81% is the cross-industry average.
- Deception. How many AI-generated outputs aren’t just inaccurate, but also misleading, factually incorrect, or fabricated? 8% is the cross-industry average.
- Decision quality. In how many business decisions based on AI-driven analysis does AI add value, according to end users? 85% is the cross-industry average.
- Latency. How quickly do AI systems produce output? 1.01 seconds is the cross-industry average.
These numbers will change. AI is improving. And companies are figuring out how to govern it in ways that both manage risks and boost business outcomes. For example, financial institutions often embed AI governance within workflows for continuous supervision, validation, and compliance—and this seems to accelerate value creation. Of all sectors surveyed, financial services reported the highest level of cost takeouts from AI transforming processes: 29%, compared to the 20% average.
Where you can create AI value faster: top-performing areas
Which areas in which functions can benefit fastest, and how can you match or surpass your competitors? Once again, benchmarks can help you learn from your peers and opportunities. In our data, success stories have a common theme: AI is embedded into workflows and designed to drive intelligent automation and higher decision quality. Average reported cost savings range from 14 to 29% depending on the industry, while the average reported improvement in business decision quality ranges from 83 to 88%. Many other benefits are evident, too.
Finance. AI initiatives in finance typically target financial close times, invoice processing and accounts payable (where AI can boost financial forecasting accuracy and speed by up to 40%), and predictive forecasting and FP&A modernization (where AI can cut cycle times for demand forecasts and scenario modeling from weeks to under a minute.)
IT. IT value from AI is concentrated in few areas: cybersecurity (where it cuts cycle times for incident resolution by 52% to 60%, depending on industry), and workload optimization and cloud operations, and application development (where it cuts cycle times in SDLC by 47% to 52%, depending on industry). Across industries, IT is consistently a top three function for AI budget allocation, in part because it’s also an AI enabler.
HR. AI investments in HR primarily flow to talent acquisition (where AI improves profile match quality and accelerates hiring throughput), predictive workforce planning, and learning and development.
Operations and supply chain. The top three targets for AI investment are demand forecasting and inventory optimization, logistics and distribution, and intelligent workflow automation. Benefits are seen in both cost takeouts and real-time responsiveness.
Marketing and customer-facing functions. AI investments here are driving content personalization and velocity (where AI cuts the idea-to-going-live time to 10 days), automating customer interactions (where industry averages range from 21% to 35%), and enhancing marketing analytics.
Make AI an enterprise advantage: 3 moves to make today
AI benchmarks can’t do it all. But they can give you a concrete way to measure your resource allocation, scale, and performance both against peers and your own goals, so you can make better informed strategic and operational choices. Here’s how to get started.
- Set your baselines. Based on data from your industry and your own capabilities and business priorities, create benchmarks across financial, operating, trust, functional, and workforce dimensions.
- Choose your owners. Assign responsibility to a small team of senior executives with “skin the game” (such as business or functional leads) to monitor and meet your AI metrics, with a goal of driving enterprise-wide value.
- Follow the numbers. Decide where to scale, or course correct or exit, based on clearly defined performance thresholds, not end user enthusiasm.
With this structured approach, based on external peer benchmarks, you can use data to more rigorously set baselines, track progress, identify gaps, and focus efforts, so you can better achieve your priorities for AI that creates measurable value at scale.
Methodology
This analysis combined PwC survey data and PwC research to benchmark enterprise AI investment and performance across major U.S. enterprises. Our 2025 AI Metric Survey captured cross-industry responses from 70 senior leaders across multiple business functions including IT, Finance, Operations, Sales & Marketing, and Strategy. Respondents predominantly represented large organizations with revenue exceeding $10 billion, and workforce sizes ranging from 1,000 to more than 30,000 employees. Our research considered data sourced from CapIQ and evaluated over multiple years from Jan 1, 2023, through December 31, 2025.
To assess the relationship between AI investment and enterprise outcomes, the analysis evaluated companies based on AI investment intensity expressed as a percentage of revenue and compared performance outcomes relative to industry peers. Companies investing above defined AI investment thresholds of 1.6% of their annual revenue were analyzed against those investing below the threshold to determine whether higher investment intensity corresponded with stronger financial outcomes. In parallel, the PwC Survey data was used to benchmark operating characteristics including investment allocation across technology, training, and transformation activities; automation penetration across business functions; AI talent intensity; and enterprise trust metrics such as inference accuracy, decision quality, and system latency. Combining these data sources enabled the benchmarking to link investment intensity and operating model patterns with measurable financial and operational performance across industries.
Measure AI performance with greater clarity
Use benchmarks to help guide AI investment and operational decisions.
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}
Contact us
Follow us
