BCG's 2026 Applied AI Index, published in September, contains a number that should reset how enterprises think about AI adoption: more than half of companies now generate measurable value from artificial intelligence. The figure is 50%, up sharply from previous years, and it is not limited to tech firms or early adopters. It spans 20 industries, from finance and pharmaceuticals to manufacturing and retail.
The more important number, for anyone buying or building agentic AI, is 22%. That is the share of total AI value now coming from agentic systems, up from 17% in 2025. BCG projects it will reach 39% by 2030. In other words, the fastest-growing part of enterprise AI is the part that acts independently.
What the index measures
BCG surveyed more than 1,300 CXOs and senior leaders across 20 industries for the 2026 edition of its Applied AI Index. Over half of participating companies reported annual revenue above USD 5 billion. The survey classifies companies into four maturity groups: future-built (the most advanced), scaling, experimenting and stagnating.
Only 7.5% of companies qualify as future-built. Another 41% are scaling, meaning they are deploying AI across multiple lines of business and seeing tangible returns. The stagnating group has shrunk to just 4.5%, down from 14% in 2025. The gap between leaders and laggards is real, and it is widening.
Future-built companies show 2.4 times the top-line growth of laggards. They reach AI value and impact in about 11 months, compared with 14 months for scaling companies and 17 months for those at the back of the pack. Speed is not a lucky break. It is the result of deliberate choices in strategy, investment and governance.
Why agentic AI is driving the shift
Agentic AI is now the fastest-growing segment of enterprise AI value because it moves beyond assisting workers to executing work. Where earlier phases of adoption focused on tools that made people faster, agentic systems are taking on end-to-end tasks with limited human intervention.
BCG's data gives three reasons why this is accelerating.
First, process redesign is now the primary source of value. Future-built companies expect 70% of AI value to come from reshaping and inventing business processes, not simply deploying better tools. That means the companies getting the most from AI are the ones rethinking how work gets done, not the ones buying the most software.
Second, the hybrid workforce is becoming the default. Fifty-five per cent of future-built companies are planning for a combined workforce of humans and autonomous AI agents. Eighty-nine per cent expect AI to generate new work, versus just 11% who expect it to replace existing jobs. The narrative of AI as a job elimination tool is being replaced by a more practical one: AI as a workforce augmentation layer that creates roles as it reshapes them.
Third, data infrastructure is becoming the price of entry. Ninety-five per cent of future-built companies are undergoing a data transformation, and more than 66% have committed to a single enterprise-wide AI platform with a common architecture and control plane. The days of departmental AI pilots running on separate data are ending. Buyers who cannot unify their data layer will struggle to deploy agents at scale, because agents need access to the same information humans do.
The spending signal
Enterprise AI spending has more than tripled, from 1% of revenue in 2025 to 3.3% today. Crucially, much of that spending sits outside the traditional IT budget, covering AI talent, governance and operational infrastructure. This is significant because it means AI investment is no longer a technology decision. It is a business decision that touches every function.
BCG also found that future-built companies manage token and compute spending deliberately. Half actively encourage adoption to maximise usage, while 22% set explicit limits or controls. The message is not that companies should spend less on AI. It is that they should spend with intention, measuring what works and stopping what does not.
Autonomy is coming faster than expected
The index reports that 42% of companies expect agents to act autonomously by 2030. That is not a speculative long-range forecast. It is a majority view among senior leaders that, within four years, a significant portion of enterprise decision-making will happen without direct human sign-off.
This expectation carries implications for governance. If autonomy is the destination, the infrastructure to monitor, audit and control agent decisions needs to be in place before the agents arrive. Retrofitting governance onto autonomous systems is significantly harder than designing it in from the start.
What this means for enterprise buyers
For buyers evaluating agentic AI platforms, the BCG index points to four practical conclusions.
First, AI maturity is no longer determined by sector. BCG found future-built companies in all 20 industries surveyed. The gap between winners and laggards is a leadership and execution gap, not a sector gap. If your competitors in a different industry are moving faster, they will pull ahead on cost and speed regardless of what sector you are in.
Second, value concentration matters. Fifty-four to 78% of AI value sits in core business functions, with technology and R&D ranking as the first or second driver across all sectors. Marketing, sales and customer service also showed measurable returns. The lesson is to focus investment in a few high-impact areas before expanding, rather than spreading AI thinly across dozens of pilots.
Third, governance is now a competitive enabler, not a checkbox. BCG found that more than half of senior AI decision-makers are bypassing their own governance processes to speed agentic deployments. That is a warning, not a recommendation. Companies that build governance into their agent platforms from day one will outpace those that retrofit it later.
Fourth, platform consolidation is under way. The most successful companies are committing to a single enterprise-wide AI platform with a common architecture, rather than stitching together multiple vendor stacks. For buyers, this means vendor selection is becoming more consequential. The platform you choose in 2026 will likely be the platform you are running in 2030.
What this means for suppliers
For companies building and selling agentic AI solutions, the index sends two clear signals.
The first is that buyers are no longer asking whether AI works. They are asking how fast it delivers value and how well it integrates with what they already have. Proof-of-concept fatigue is real. Suppliers that can demonstrate time-to-value in months, not years, with clear metrics and named customers, will win procurement cycles.
The second is that data readiness is the new gate. If 95% of AI leaders are investing in data transformation, then any agent platform that does not help buyers clean, unify and govern their data is solving the wrong problem. The best agent in the world cannot generate value if the data it sees is fragmented, stale or inaccessible.
The Agentic Expo angle
Agentic Expo, at Olympia London on 23 to 24 March 2027, exists because the shift BCG documents is not theoretical. It is happening now. The buyers who attend are not exploring whether agentic AI is real. They are deciding which platforms, governance frameworks and partners will carry them from where they are today to where they need to be by 2030.
The 22% figure is the headline. But the underlying story is faster value, higher spending, broader autonomy and sharper focus. Companies that have solved process redesign, data infrastructure and governance are pulling away from those that have not. Agentic Expo is where that gap gets closed.
Sources: BCG, "The Formula for Agentic AI Value", September 2026; ZDNET, "50% of companies generate value from AI now — here's how", 13 October 2026.