Artificial intelligence opens up a new world of opportunity and complexity for CIOs and their organizations. As they push for growth and efficiency, tech leaders are playing a bigger role in business transformation. They face dual challenges: do more with less, while securing the resources to pilot and scale AI. Turning AI’s promise into performance, without letting costs and complexity spiral, is becoming essential to staying ahead.

As part of that drive, many companies are already using AI to help manage costs. For example, GPTs that classify spending data, tagging invoices and mapping them to general ledger entries, can help identify shadow spending and rein in technology costs outside the CIO’s control. Other uses include assessing applications for duplicate functionality or low usage, making it easier to retire less valuable software and reducing related costs by up to 30%.

But AI is also creating complexity and new costs: In a recent Bain survey of more than 400 tech leaders, 69% expected to see more than a 5% increase in spending on AI (see Figure 1). Surging business demand for AI is fueling rapid technological change and major process overhauls. At the same time, underlying AI technologies are changing quickly, with replacement lifespans measured in months, not years.

Beyond direct costs, AI increases the complexity of running a business and supporting the technology’s ever-increasing pace of change, the necessary retooling of architecture, and the development of new operating models. Enterprises are layering AI models, agents, and platforms onto already fragmented digital ecosystems and aging core systems. This creates new integration challenges and, in some cases, higher operating costs (see Figure 2). It also requires more data collection and analysis, higher data storage costs, and new guardrails to track decisions and improve efficacy of autonomous agents and systems acting with little to no human input. Some AI-enabled workloads, especially those powered by large language models, may end up being significantly more expensive than the traditional technologies they replace—at least in the near term.

Given these added cost pressures, it’s more important than ever for CIOs and other tech leaders to capture savings where they can.

Where AI reduces operating costs

AI is built to take on complex, knowledge-intensive work and perform it faster and more efficiently than ever before. Enterprise technology is a prime candidate, with real opportunities to streamline operations, cut waste, and reduce costs significantly.

• Smarter tech spending. AI brings near real-time visibility into IT costs, automatically tagging expenses and surfacing hidden spending like shadow IT. That makes it easier to flag overspending and tighten control, even beyond the CIO’s line of sight. A global media company used AI this way, consolidating data from more than 80 general ledgers and identifying tens of millions of dollars in shadow IT spending. The improved transparency helped the company classify and benchmark costs, enhance oversight, and implement targeted controls, which together created more savings and improved stewardship of IT resources.
• Sharper view of resource use. AI-powered usage analytics help teams see exactly what’s being consumed and where. That means more accurate forecasts, less waste, and smarter infrastructure decisions, so the business only pays for what it actually needs. A global life sciences company used AI to spot where they were overspending on cloud services, such as servers running when they weren’t needed or storage they no longer used. With a clearer view, they were able to scale their environment to fit actual needs, cut unnecessary spending, and put smarter controls in place. This delivered sustained savings and laid a foundation for future improvements.
• Rationalizing application portfolios. AI can spot overlap and underused tools across the application stack, helping teams retire redundant software. That kind of pruning typically cuts software and maintenance costs by 10% to 30% and streamlines how tech gets managed. A specialty chemicals company used AI to scan its application inventory to flag duplicate and underused software, map the inventory against software costs, and identify about a quarter of the portfolio and nearly 30% of spending as no-regret opportunities for consolidation or retirement.
• Running with AI ops. With AI built into operations, teams can predict and prevent incidents, automate fixes, quiet the noise from false alerts, and reduce alter fatigue—the exhaustion that results from too much rapid change. It’s a shift toward self-healing systems—reducing outages, cutting manual effort, and simplifying complexity. For example, a content-management software provider used an AI tool to detect anomalies and alert teams before problems fully developed. This helped engineering teams respond quicker, which reduced the company’s resolution times by 15%.
• Speeding up software delivery. Generative AI coding assistants and automated testing tools help development teams move faster and smarter: generating code, refactoring legacy systems, writing tests, reviewing code, and resolving bugs faster and with higher consistency. The result? Software development cycles shrink by 20% to 30%, with lower labor costs, better quality, and faster time-to-market.

How AI controls business demand for technology

Just as AI is helping IT teams get smarter about how they spend, it can also help the broader enterprise manage the demand for technology more effectively. IT leaders are more central than ever to underwriting transformations, and they should be using AI to transform more efficiently, embedding AI into operations, and optimizing talent and operating models.

• Rewiring the transformation life cycle. Reducing IT costs means looking beyond just software development. Between 50% and 65% of the work in a tech transformation is administrative (analysis, design, change management) rather than hands-on-keyboard development work. AI can streamline every phase, from upfront research to automated design-to-code workflows. To unlock maximum benefit, cost management should extend across the entire life cycle.
• Embedding AI in operations. AI is changing technology support models with conversational customer support agents and automated ticket resolution. When AI is embedded into daily operations, it cuts costs across model maintenance, data management, and vendor management. For example, an airline that deployed AI and other automation to help customer support agents reported a 40% increase in productivity.
• Flexing the workforce and partner ecosystem. Large-scale transformations rely on a mix of internal teams and external partners. AI helps orchestrate that ecosystem—matching the right work to the right resources, automating governance, and supporting better decisions. The result: higher-quality execution with greater efficiency.

AI is also reshaping the delivery of IT services. Some service providers like Globant are moving to AI-driven, outcome-based pricing models. Internal tech teams will also need to rethink how they source, manage, and deliver IT services, and most will need to develop new roles, workflows, and ways of working that embed AI into day-to-day operations. (For more on this, read the Bain brief “AI Pods as a Service: Modular, Scalable, and Built for Speed.”)

Controlling costs while scaling AI

AI has the potential to reduce overall technology spend, even as demand for digital grows. But impact only comes with disciplined scale. IT and transformation teams must be deliberate in how they deploy AI across the enterprise:

• Fund AI with AI. While investing in AI is essential, it can also pay for itself. By using AI to streamline operations and reduce tech costs, IT can help offset the expense of broader AI adoption, creating a flywheel that funds transformation with its own efficiencies.
• Simplify with discipline. As AI agents proliferate, strong architecture governance becomes non-negotiable. Think lean: simplify architecture choices, set clear standards, and build in controlled environments. One powerful example is a tiered AI model strategy—using smaller, fine-tuned models for high-volume, routine tasks at a fraction of the cost, while reserving large models for complex, high-impact uses. A new supply chain is emerging—from optimized chips to multi-model platforms—designed to help enterprises operate on cost-efficient AI architectures.
• Embed AI in the operating model. The real value of AI shows up when it’s woven into how the business runs, from delivery to governance. Companies that have already shifted to cloud-based infrastructure tend to be more effective at managing AI costs. Scaling AI across both IT and business operations means rethinking tooling, updating governance and controls, and cultivating an AI-first mindset across teams.

Executives know they can’t optimize what they don’t understand. Better transparency of technology costs can help them make sure that spending is focused on strategic priorities. Following through with disciplined cost management ensures that these investments deliver the expected returns. AI can help with both.