Welcome to our monthly rundown of what’s happening in AI for October 2026.
Is enterprise AI paying off yet? The research published over the past month says yes for about half of companies, though the gains are modest for most of them.
In other news, agents are moving into production faster than the controls meant to govern them. AI costs have become a big focus for CFOs. And while most companies know that AI will fundamentally change people’s roles, very few have a plan for how to adapt.
Underpinning most of these findings is a measurement problem. Many companies never decided what success or failure would look like, so they can’t tell which projects to scale and which to shut down.
Let’s get started.
Is AI successful in most companies? It depends on who you ask
BCG’s Applied AI Index, a survey of 1,330 executives published September 30, found that nearly half of companies now capture value from AI, up from 5% in BCG’s 2025 research. Companies BCG classifies as “future-built” show 2.4 times the revenue growth and 2.8 times the EBITDA growth of laggards. KPMG’s Q3 AI Pulse, released September 24, found that nearly six in ten leaders at companies with more than $1 billion in revenue report measurable value. The most common gains are productivity (55%) and faster decision-making (49%). But only about a third (37%) cite financial gains.
Gartner’s survey of 2,002 CIOs, released October 19, sets a higher bar. Only 13% of respondents say their enterprise has realized significant or transformational value from AI, and Gartner puts use-case success at roughly one in five. Read together, the surveys show that most companies see AI working somewhere in the business. The number shrinks as the bar rises. About a third report financial gains, and only one in eight CIOs report significant or transformational value.
Any business decision has to come back to how it improves the company’s revenue and profit. Unfortunately, too many execs can’t quantify P&L impact for their AI initiatives. That has to change.
Finance is watching AI costs
AI spending has roughly doubled as a share of revenue, from about 1.7% in late 2025 to 3.3%, according to BCG, and more than 80% of it now sits outside the IT budget. Gartner found that 58% of CIOs face more pressure to deliver AI cost improvements and 51% expect AI to raise the total cost of ownership of technology. Gartner traces part of the problem to pricing. AI costs are set at runtime and vary with usage, so they’re harder to forecast than traditional software.
Companies are responding with tighter controls. KPMG found that 74% now include cost reviews in AI approvals, up from 61% a quarter earlier, and 43% have set usage or token budgets. Forrester expects B2C marketing leaders to cut about a third of their AI initiatives in 2027 because usage-based pricing turns every agent interaction into a billable event. A project with no defined return has a hard time justifying a variable bill, which gives companies one more reason to set success criteria before launch.
Too many companies get a shock when the monthly token bill arrives. They want better demand forecasting, proactive efforts to reduce consumption, and costs that are consistent and predictable.
Was success ever defined?
The same Gartner survey we mentioned earlier points to a key problem with many AI projects. Only 31% of CIOs say they have specific criteria for moving an AI pilot into production. Just 17% report having guidelines for retiring AI that isn’t delivering. Organizations with production criteria are 2.3 times more likely to be among Gartner’s top performers, and those with retirement guidelines are 2.6 times more likely. Without either, pilots linger, low-value projects keep drawing budget, and no one can say whether a solution moved the business forward.
BCG’s study of nearly 400 revenue leaders helps explain why business and IT leaders may see success that doesn’t show up in financial results. Of the organizations surveyed, 81% have AI adoption goals, but only 36% set formal targets and track them. In the absence of those targets, reported gains follow what’s easy to count. A third of respondents report less administrative work, while 13% report better win rates and 8% report higher quota attainment.
Those changes have financial value, but you need concrete ways of measuring it, agreed before you sign a contract.
The bots are coming, and they want hard numbers
Forrester predicts that AI agents will influence 75% of B2B purchase decisions in 2027 and that two-thirds of solution providers will lose deals because buyers can’t verify their claims. Got no proof? You’ll get no sale. In BCG’s study, 51% of sales leaders say customers now expect a customized proof of value, including ROI evidence, yet just 32% say AI is helping them build that proof. A company that never set targets for its own AI will struggle to show verified results to a buyer’s research agent or CFO.
Agents are good at ignoring marketing hyperbole and spin, often better than people are. You need rock-solid proof, and you can’t get that without baseline information and a clear way to quantify results.
Agents are outrunning governance
Gartner reports that 71% of organizations expect to deploy agentic AI within 12 months, while only 21% have agentic AI governance in place. Gartner’s IT Symposium keynote reported that AI risk is growing faster than AI value. As Kristin Moyer, VP, Distinguished Analyst and Gartner Fellow, said:
“That imbalance makes it difficult to create value faster than AI creates risk. Managing value and risk is about building the right kinds of momentum while avoiding the wrong kinds of drag.”
BCG found that 42% of companies expect to give agents real decision-making authority by 2030, but just 5% have the full set of controls in place today. Interestingly, KPMG’s numbers show confidence rising as controls slip. The share of leaders confident in their AI governance rose from 57% to 73% in a single quarter. However, over two quarters, the share of leaders reporting that their companies are building controls directly into their agents fell from 43% to 30%. Meanwhile, multi-agent system development jumped from 6% to 25%, which means more decisions passing between agents with no human review. That’s a sobering picture.
On the plus side, some companies are drawing firm lines. Nearly half of KPMG’s respondents have defined high-risk use cases where agents can’t decide on their own, and 75% of banking leaders keep agents away from sensitive data without human oversight.
Preventing catastrophic outcomes is important, but governance is essential for all agentic solutions. Safeguards must be part of every AI project from the first step.
Workforce plans lag
Only 4% of CIOs in Gartner’s survey have explicit plans to redesign roles for a workforce that combines people and agents. Those that do are 2.4 times more likely to be top performers. BCG’s respondents expect workforce reductions of 10% to 15% by 2030, with most of the pressure on middle management. They also expect dedicated AI roles to grow from 7% of the workforce to 22%.
A new McKinsey Global Institute report estimates that AI and automation will cut demand for about 36 million US jobs by 2035 and create demand for about 41 million jobs. Roughly 11 million workers could need to leave their occupations entirely. Lower-wage workers are 7.6 times as likely as higher-wage workers to need a new occupation, and only one in seven displaced workers has a direct path into a growing job.
Skills are at risk even in companies that keep their people. Half the executives surveyed by the BCG Institute say they already see AI-related deskilling, which BCG’s David Martin attributes to over-reliance on AI summaries and to handing strategic work to AI.
Getting value from AI depends on people with the skills to direct and check it. With only 4% of CIOs planning for that workforce, most companies will have trouble finding both the talent and the returns.
How RapidCanvas helps
Each of this month’s findings comes back to setting goals and limits before an AI project scales. At RapidCanvas, we use a Hybrid Approach™, combining human experts with a proven agentic AI platform to deliver a solution that’s customized to your goals, workflows, and tech stack.
Our platform is built on three connected products. The Enterprise Context Engine™ captures and curates your business knowledge, and Visual Development Studio builds solutions from reusable capabilities. The Operational Harness governs execution in production, so agents act within the limits you set. Together they give you a foundation you understand and own, with knowledge and capabilities that compound across engagements.
If you’d like to discuss how to set measurable goals for your next AI initiative with an expert, RapidCanvas would love to help. Contact us to schedule a time, or visit our website to browse dozens of case studies and read verified customer reviews on G2.
News from RapidCanvas
It’s been a big month for us. We announced a $29 million Series B round and our acquisition of Entytle, which brings AI-driven execution to the industrial aftermarket.
Our $29M Series B
On October 16, we announced a $29 million Series B round led by Vertex Ventures SEA & India, with Hitachi Ventures joining as a new investor. Existing investors Accel, Peak XV Partners, Valley Capital Partners, and Titanium Ventures also participated, bringing our total funding to $52.5 million.
The funding will expand our go-to-market teams and our network of private equity and advisory partners. It will also grow our field AI engineering team in the US, Brazil, and India and support continued investment in our three products.
Our annual recurring revenue has grown 8x over the past six quarters, and more than 40% of our customers run multiple AI use cases in production. So far in 2026, more than 75 enterprises have made 1.4 million AI-led decisions with applications built on our platform.
Our acquisition of Entytle
On October 28, we announced our acquisition of Entytle, the company that pioneered the Autonomous Aftermarket for industrial OEMs.
Manufacturers sell machines that run for 10 to 30 years and generate parts and service revenue long after the sale. Most OEMs don’t have enough people to stay in touch with every customer and asset, so they cover their top accounts and leave the rest untended. Entytle’s AI agents monitor activity across an OEM’s installed base, recommend sales, marketing, and service actions, and handle prospecting, quoting, and customer service outreach in the background.
The acquisition brings Entytle’s technology, employees, and outstanding customers to RapidCanvas. Those customers include Peerless Pump, Baker Hughes, GEA, and Veolia.
That’s it for this month. Thanks for reading. Subscribe to our newsletter for more curated AI insights from leading business publications.
