Enterprise AI is moving beyond experimentation, with nearly half of companies now generating value from AI, according to Boston Consulting Group’s 2026 Applied AI Index. But as spending rises and companies prepare to give AI agents greater autonomy, governance and control systems are emerging as the next major challenge.
The enterprise AI market is entering a different phase. Instead of asking whether artificial intelligence can generate business value, companies are increasingly dealing with a harder operational question: how to scale that value while keeping increasingly autonomous systems under control.
That is the central finding of Boston Consulting Group’s Applied AI Index 2026, based on a survey of more than 1,300 C-suite executives and senior leaders across more than 20 sectors. BCG found that almost half of companies are now generating value from AI, a substantial change from its 2025 research, which identified only 5% as generating significant value.
The shift is not limited to a small group of AI leaders. BCG classifies 7.5% of companies as “future-built” and another 41% as “scaling”, meaning they are already generating some value and expanding their AI deployments. Together, those groups account for nearly half of the companies in the study.
The most mature organizations are also reporting stronger business performance. BCG says future-built companies generate 2.4 times the top-line growth of laggards, while the most advanced companies have also recorded substantially higher shareholder returns and EBITDA growth.
That matters because enterprise AI spending is increasing rapidly alongside adoption. BCG’s research puts AI spending at 3.3% of company revenue, more than double the level reported in late 2025, with 80% of AI spending now outside enterprise IT budgets.
The change effectively turns AI into a business transformation investment rather than a conventional IT project. Marketing, operations, customer service, finance and other functions are increasingly funding and deploying AI directly.
But the expansion is also exposing a governance gap.
BCG found that 42% of companies expect AI agents to operate autonomously by 2030, while only 5% currently have the relevant controls in place. Agentic AI already accounts for 22% of AI value in the 2026 sample, up from 17% in the previous year’s research, and BCG projects that share could reach 39% by 2030.
The implications are significant for AI platforms and infrastructure providers. An enterprise deploying an AI agent to summarize documents or generate code can maintain relatively tight human oversight. An agent that can approve transactions, alter records, initiate procurement or make operational decisions requires a different control architecture.
BCG identifies oversight, rollback mechanisms, security, auditing and cost controls as part of the infrastructure required for greater agent autonomy. Its research also found that organizations with all six identified AI controls generate three times as much agentic AI value as companies with only one control.
This is pushing AI governance toward becoming an infrastructure layer, rather than a compliance process added after deployment. Identity, permissions, model access, observability, policy enforcement and audit trails increasingly need to work across agents and AI applications.
BCG’s broader research supports that direction. Its August analysis of enterprise AI control planes argues that organizations deploying agents across multiple platforms face fragmented security, cost and operational oversight, increasing the need for centralized governance.
The workforce impact is also shifting from simple automation toward organizational redesign. BCG’s latest index says companies expect AI to reduce workforce requirements by roughly 10% to 15% by 2030, while 89% expect AI to generate new work. The report says 55% of future-built companies already practice strategic workforce planning for an agentic environment, compared with 17% of laggards.
That suggests the next stage of enterprise AI will not be determined solely by model performance. Companies will need to redesign workflows, establish accountability and determine which decisions remain human-controlled and which can be delegated to software.
The emerging technology stack reflects that change. AI development frameworks and model gateways provide access to increasingly capable models, while agent platforms add orchestration, memory and tool access. Governance systems then determine what those agents can access, what they are permitted to do and when human intervention is required.
For companies still evaluating AI, BCG’s findings also complicate the idea that the market has simply split between AI winners and everyone else. Nearly half of surveyed organizations are now generating value, but a similarly large group remains in earlier stages of maturity. The difference increasingly appears to be how companies connect AI investments to business outcomes, redesign processes and build the infrastructure required to scale.
The result is a new enterprise AI equation: models provide capability, platforms provide execution, and governance determines how far that capability can safely operate.
As agents move from assistants toward autonomous decision-makers, that final layer may become one of the most important components of enterprise AI infrastructure.
Market Landscape
Enterprise AI is moving from isolated pilots toward scaled deployments, but maturity remains uneven. BCG’s 2026 research identifies 7.5% of surveyed companies as future-built, 41% as scaling, 47% as emerging and 4.5% as stagnating.
Agentic AI is accelerating this transition. BCG estimates that agentic systems already account for 22% of AI value and could reach 39% by 2030. Yet only 5% of surveyed organizations currently have the full set of controls needed for autonomous agents.
This creates a growing market for AI control planes, model gateways, agent orchestration, observability, security and governance platforms. The competitive landscape now includes hyperscalers such as Microsoft, Google and Amazon, enterprise software providers such as Salesforce and specialized AI infrastructure companies.
The central technology challenge is shifting from model availability toward reliable, measurable and governable AI deployment.
Top Insights
- Nearly half of surveyed companies now generate value from AI, indicating enterprise adoption is moving beyond isolated experimentation and into broader operational deployment.
- AI spending has reached 3.3% of revenue, with 80% occurring outside enterprise IT, turning AI into a company-wide business investment.
- Forty-two percent of companies expect autonomous AI agents by 2030, but only 5% currently have the necessary controls in place.
- Agentic AI already represents 22% of AI value in BCG’s 2026 sample and could reach 39% by 2030.
- Enterprise AI maturity increasingly depends on workflow redesign, financial measurement, workforce planning and governance rather than model capability alone.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
