AI Leaders Achieve 4x ROI as Governance and Skills Drive Adoption

Companies leading in artificial intelligence adoption report an average return of four times their AI investment, compared with 2.7 times among slower adopters, according to IDC research commissioned by Microsoft.

The study surveyed more than 5,200 senior IT and business leaders across 28 industries, examining how organizations translate AI investment into measurable business outcomes. It identified a substantial gap between companies achieving results from individual AI projects and those deploying the technology successfully across their operations.

IDC classified the top 10% of organizations on its AI maturity scale as Frontier Firms, while the bottom 30% were categorized as slow adopters. Among Frontier Firms, 39% reported AI returns of at least five times their investment, compared with just 4% of slower adopters.

• Frontier Firms report average AI returns of 4x, compared with 2.7x among slow adopters, and demonstrate stronger capabilities in measuring investment performance.
• Some 96% of Frontier Firms are deploying AI agents at scale, while 87% have agent governance processes, compared with 51% of slower adopters.
• Only 20% of surveyed organizations have comprehensive AI training strategies, despite 67% reporting high workforce trust in the technology.

Although 66% of business leaders report substantial or transformative AI results, only 48% have scaled that value across multiple business units.

Enterprise AI adoption expands faster than measurable returns

AI is now used across an average of 6.5 business functions within surveyed organizations, with IT infrastructure, cybersecurity and customer service among the leading applications.

Industry-specific deployment is growing particularly quickly as companies extend AI beyond general productivity tools.

The business benefits being pursued have also diversified. Employee experience, revenue growth, customer experience, innovation and risk reduction now rank alongside cost savings as important sources of value.

Organizations increasingly assess AI platforms using security, accuracy, compliance, scalability and cost rather than focusing exclusively on efficiency.

Only 61% of surveyed organizations are highly confident in their ability to measure AI business value and return on investment, compared with 93% of Frontier Firms.

The research indicates that deploying AI does not necessarily translate into consistent financial returns across an organization.

Companies may achieve results in individual departments while lacking the systems needed to measure performance, replicate successful applications or identify where further investment is justified.

IDC research stated organizations achieving value at scale manage investments, models, agents, governance and employee skills as interconnected business capabilities.

The reported return differences represent associations within the survey rather than evidence that any single AI practice directly causes higher profitability.

Frontier Firms deploy AI agents across business operations

AI agent adoption is a major distinction between organizations at different maturity levels. Among Frontier Firms, 96% report rolling out agents at scale, while 53% orchestrate agents across workflows. Across the wider surveyed population, 36% of organizations remain in the pilot stage of agent deployment.

Frontier Firms also report a 3.8x return on agent investments, compared with 3.0x among slow adopters.

Nine out of ten Frontier Firms spent more on AI than originally planned, combining higher investment with broader deployment across business functions.

The research identifies five common practices among organizations reporting the strongest AI results: investing ahead of initial budgets, deploying AI across multiple functions, scaling agents, establishing governance and developing workforce capabilities. These companies generally treat AI adoption as a change to business processes rather than an isolated technology implementation.

Microsoft cited Bristol Myers Squibb as an example of workflow-based deployment. Its RABIT agent supports project managers by connecting information across meetings and reducing administrative follow-up. Employees using the system save up to an estimated 20% of their time.

AI governance expands but security confidence remains limited

Agent governance has become increasingly important as AI systems move beyond generating content and begin executing actions within enterprise workflows.

IDC found that 65% of surveyed organizations have established processes for governing AI agents. However, only 60% report high confidence in their ability to address AI security and governance risks.

Among Frontier Firms, 87% have formal agent oversight procedures, compared with 51% of slow adopters.

The gap between governance adoption and security confidence suggests that formal controls do not automatically translate into effective management of AI-related risks.

Organizations deploying autonomous or semi-autonomous agents must account for access permissions, data security, decision-making boundaries and human oversight (see how AI Reshapes Workforce).

Governance requirements therefore extend beyond evaluating model outputs to controlling how agents access systems, process sensitive information and perform business activities.

Microsoft highlighted Conagra Brands, which established an AI governance committee to review risk factors, identity management and access controls before expanding employee adoption. The company also introduced observability and responsible AI practices as part of its implementation framework.

Workforce training remains the largest gap in AI adoption

Employee confidence in AI is significantly more widespread than formal workforce development. The survey found that 67% of organizations report high levels of employee trust in AI, yet only 20% have comprehensive, organization-wide training strategies.

The difference is particularly pronounced between mature and less mature adopters.

Approximately 80% of Frontier Firms have an enterprise-wide AI skilling strategy, compared with only 3% of slow adopters.

The findings suggest that technological deployment is advancing faster than organizational preparation.

Effective adoption requires employees to understand not only how to use AI applications but also how to redesign existing processes, evaluate outputs and determine which activities should remain under human control.

Kohler Chief Digital Officer Eric Tabor emphasized the importance of business-focused training, including the ability to analyze processes and convert them into AI-enabled workflows.

Engineering and consulting company Black & Veatch reported 98% adoption of Microsoft 365 Copilot following a structured change-management program.

The company used early-adopter groups to identify practical business applications and support wider implementation.

AI platform selection balances security, performance and cost

The research also examined how organizations select AI technology providers. Trust and security, access to advanced technology, and model performance were equally important selection criteria, each cited by 44% of respondents.

Companies must balance these requirements against infrastructure expenses, intellectual property protection and the sensitivity of business data.

Using different AI models for different activities can allow organizations to match performance requirements with appropriate levels of cost and security.

The wider challenge is integrating those models into operational processes where agents can perform useful tasks without introducing unacceptable risk.

Business leaders identified creativity, innovation and business reinvention as areas where AI could expand employees’ capabilities, cited by 41% of respondents.

Customer-facing activities and relationship management were identified by 39%, alongside analytical work and data interpretation at the same level.

AI budgets expected to increase as companies focus on ROI

Enterprise AI spending is expected to continue growing despite persistent challenges in measurement and governance.

The IDC survey found that 71% of organizations expect their AI budgets to increase by at least 10% over the next 12 months.

Higher spending is likely to place additional pressure on companies to connect AI investment with measurable operational and financial performance.

More than seven in ten organizations expect double-digit AI budget growth, while nearly four in ten are still piloting AI agents.

The findings indicate that successful implementation depends on establishing performance baselines, tracking results before and after deployment and directing investment toward applications with demonstrated business value.

For organizations expanding agent deployment, these measurements also need to account for security, compliance, workforce productivity and the cost of ongoing supervision.

Microsoft presented the findings ahead of its Ignite conference on November 18, 2026, where IDC planned further discussions on differences in AI returns between leading organizations and slower adopters. The research identifies governance, measurement and workforce training as the principal organizational capabilities distinguishing companies that scale AI from those achieving results primarily through individual projects.

………………

EXPERTS: Eric Tabor – Kohler Chief Digital Officer, Kathy Lange – IDC research director, Alysa Taylor – Microsoft Chief Marketing Officer, Commercial Cloud & AI

by Peter Sonner, Lead Tech Editor at Beinsure