By Kathy Gibson – Artificial intelligence (AI) is starting to pay off for early enterprise adopters. More than half of companies that have implemented AI are now seeing a return on investment, with the most value reported by future-built organisations.
This is the headline finding from the Applied AI Index 2026, launched today.
Nicolas De Bellefonds, MD and senior partner of BCG X, points out that investment in AI has doubled over the last year – and 80% of this investment now sits outside of the IT organisations.
The companies leveraging agentic AI, and leading more with technology, are seeing substantially more value than others, he adds. Leaders see three-times more value than laggards, and future-built companies are seeing a significant 6-times more.
However, control is emerging as a vital area, Bellefonds says. Today, just 5% of companies have implemented controls, although there is a move to close this gap.
The impact on the workforce is going to be significant, he adds. While agentic AI is reshaping rather than replacing workers, organisations are still expecting to reduce workforces by 10% to 15%.
AI maturity overall is rising, but there is still a way to go. In 2026, there 51% of organisations can be described as laggards, compared to 71% in 2024. At the top end of the scale, future built leaders have risen from 4% to 7,5%.
There is no particular industry leading in AI deployment, Bellefonds says. What they do with AI seems to matter more than the particular industry.
As AI starts to demonstrate value, investments continue to rise, growing from an average of 1% of revenue in 2025 to 3,5% in 2026. In fact, AI investment is now bigger than the average IT spend, with 80% of that made outside of IT. “Business beyond IT has started to seize the AI wave and invest behind it,” Bellefonds says.
Token cost is a real concern, and seven out of 10 leaders now actively manage token cost, with 70% taking active decisions on cost: encouraging use, capping it or pulling back access. Just 30% monitor or allow token use without ever setting a limit.
Vlad Lukic, global leader: tech and digital advantage at BCG X, drills down into the ROI from AI, explaining that 79% of leaders can point to value in their financials or clear KPIs; 79% track AI value in their profit and loss (P&L) or through proxy measure such as cycle time; and 21% judge it qualitatively, with a direction sense of value.
The majority of leaders deliver AI value within a year.
Lukic points out that future-built organisations deliver value on 69% of their AI projects within the first year, with just 30% needing more than one year. Laggards, on the other hand, see only half the portfolio delivering value within a year.
Agentic AI is rapidly delivering value, with two-fifths of all AI value expected to come from agentic AI by 2030.
Lukic adds that there are six key agentic controls required to ensure safe agent autonomy: memory and data access; human in the loop; testing and shut off; interface control, clear ownership; and security and audit.
While 42% of organisations expect to be running autonomous agents by 2023, just 5% of them have the relevant controls in place today.
Jeff Walters, tech and digital advantage lead: APAC at BCG X, expands on the workforce impacts expected from AI deployment, pointing out that agentic AI will reshape more work than it replaces – but there will be job losses.
Indeed, companies expect a workforce reduction of 10% to 15% by 2030.
The job profiles likely to be reduced by AI are project management program roles (11% to 15%), middle management frontline leaders (10% to 13%), functional leadership (7% to 9%) and senior leadership (5% to 6%).
At the same time, more AI roles will be created, Walters says, with three-times more employees dedicated to AI roles – up from 7% in 2026 to 22% in 2030.
He stresses that these outcomes are what companies expect as of today, and could change. “Everyone is on a journey regarding what AI means for the workforce.”
In the meantime, seven in 10 leaders are retraining staff, with just 6% taking no action in reskilling. Forty percent of organisations are doing strategic workforce planning around AI; and 41% are revisiting how AI talent is sourced.
“Companies delivering value realise they are reshaping workflows,” Walters says. “So they are changing the nature of jobs and have to rethink pyramids, how many people are needed, and in what functions. So we are not surprised that 40% of leaders are strategically rethinking their workforces.”