Where AI Is Actually Paying Off in 2026, and Where It Isn't

As 2027 budgets close, most boards are asking a version of the same question: we have spent on AI for three years; what did we get? The honest answer, according to this year's research, is that it depends heavily on what was measured, and where the money went.

The spending is no longer marginal. BCG's Applied AI Index 2026, a survey of 1,330 executives published in September 2026, reports that AI spending has risen to about 3.3% of revenue, and that more than 80% of it now sits outside the IT budget. At that level, AI is an operating investment that competes with headcount, marketing and capital projects, and it should be judged like one.

This article reads the major 2026 studies side by side, explains why they appear to disagree, and draws out the pattern that matters for budget decisions: returns concentrate in a minority of companies that put AI into core workflows and measure financial outcomes.

The short answer

Is AI paying off for companies in 2026? For a growing minority, clearly yes; for most, not yet in the P&L. McKinsey's 2026 survey finds 37% of organizations attribute any EBIT impact to AI and about 6% get 5% or more of EBIT from it. PwC's 2026 CEO Survey finds 56% of CEOs report no significant financial benefit. BCG finds nearly half of companies now generate meaningful value. Returns are strongest where AI is embedded in core functions (sales and marketing, operations, supply chain) and workflows are redesigned.

Why the studies seem to disagree

The headline numbers range so widely that executives can find support for any position. The differences are mostly about what each study counts as "value."

Read from top to bottom, the bar rises from "we see a return" to "it shows up in earnings." Wharton's data adds a second effect: 81% of VP-and-above leaders believe ROI is positive, versus 69% of mid-level managers. Optimism falls the closer one gets to the work, and the closer the metric gets to the P&L.

The practical lesson: before debating whether AI is working, agree on which rung of that ladder your company is measuring.

Where the returns are showing up

Core functions, not the IT department

BCG's 2025 Widening AI Value Gap study estimated that about 70% of AI's potential value sits in core functions such as sales and marketing, manufacturing, supply chain and pricing, with IT accounting for about 13%. McKinsey's 2026 survey points the same way: respondents report cost reductions most often in supply chain, service operations and manufacturing, and revenue gains most often in marketing and sales, followed by product development and software engineering.

Software engineering, increasingly as a substitute for purchased software

McKinsey also found that 32% of respondents have dropped a planned software purchase because they could build the capability with agentic coding tools, and roughly one in five are scaling coding agents. For leadership teams, the implication reaches beyond developer productivity: it changes the build-versus-buy calculation for internal tools.

Companies that integrate AI into what customers buy

PwC found that companies applying AI widely to products, services and customer experience reported margins nearly 4 percentage points higher. CEOs reporting both cost and revenue gains were two to three times more likely to have embedded AI extensively.

Where money is still being burned

• Broad tool rollouts without workflow change. McKinsey's high performers are distinguished less by spending than by redesign: nearly three-quarters fundamentally redesigned workflows, versus about a quarter of other companies.
• Agents without controls. BCG reports that 42% of companies expect to give agents real autonomy by 2030, but only 5% have the full set of controls in place. Deloitte finds only one in five has a mature governance model for autonomous agents.
• Run costs nobody budgeted. About 20% of McKinsey's 2026 respondents say token and operating costs have constrained their AI use.

What the leaders do differently

BCG's 2026 index puts "future-built" companies at 7.5% of the sample, up from 5% in 2025, with another 41% scaling. Compared with laggards, future-built companies report 2.4 times the revenue growth, 2.8 times the EBITDA growth and 2.3 times the total shareholder return. These are correlations from a survey, not proof that AI caused the gap, but the operating patterns behind them are consistent across studies:

1. They concentrate. Fewer use cases, chosen for size of financial impact in core functions.
2. They redesign work. AI changes the process, not just the tool on top of it.
3. They measure money. Success is defined in cost, revenue or margin, with a baseline.
4. They invest in foundations. PwC found CEOs with strong foundations, such as responsible AI practices and enterprise-wide integration, three times more likely to report meaningful returns.

A framework for 2027 AI budgets

Sort each proposed AI investment into one of four boxes before funding it.

Editorial framework by DaCodes, built on the BCG and McKinsey findings above.

A perspective that challenges the consensus

The common narrative is that most companies are failing at AI. The 2026 data supports a more useful reading: many companies are succeeding at adoption and failing at measurement and redesign. Usage is near universal (Stanford HAI's 2026 AI Index reports organizational AI use rising from 55% in 2023 to 88% in 2025), so the gap is not access to technology. It is whether the operating model changed. That is a management problem, and therefore a solvable one.

What this means for business leaders

1. The value gap is widening, not closing. The share of high performers is small and stable in McKinsey's data, while leaders in BCG's sample report multiples of growth over laggards.
2. AI belongs in the operating budget conversation. At over 3% of revenue and mostly outside IT, AI spending needs the same business-case discipline as any operating investment.
3. Core functions deserve the next dollar. Evidence consistently points to sales, marketing, service operations and supply chain as the places where AI shows up in the P&L.

Your next 90 days

1. Days 1–30: Inventory AI spending across all budgets, not just IT. Classify each line using the 2x2 above.
2. Days 31–60: For the top two core-function opportunities, set a financial baseline and design the redesigned workflow before any build.
3. Days 61–90: Shift budget from capped productivity tools toward those two initiatives, with quarterly reviews against the financial metric.

Frequently asked questions

What percentage of companies get ROI from AI? It depends on the measure. In 2025–2026 studies, about 75% of US enterprise leaders report positive ROI (Wharton/GBK), 37% report any EBIT impact (McKinsey), and 12% of CEOs report both cost and revenue benefits (PwC).

Which business functions get the most value from AI? McKinsey's 2026 survey reports cost reductions most often in supply chain, service operations and manufacturing, and revenue gains most often in marketing and sales. BCG estimates about 70% of AI's potential value sits in core functions.

How much are companies spending on AI? BCG's Applied AI Index 2026 reports AI spending at about 3.3% of revenue, with more than 80% of it outside the IT budget.

What do AI high performers do differently? They redesign workflows (nearly three-quarters of McKinsey's high performers, versus about a quarter of others), focus on core functions, measure financial outcomes and invest in data and governance foundations.

Why do AI ROI studies report such different results? They measure different things: self-reported ROI, productivity, EBIT impact or CEO-reported financial benefit. They also survey different respondents, from mid-level managers to CEOs.

Conclusion

The question for 2027 is not whether to keep investing in AI; virtually every company will. It is whether the next dollar goes to another tool layered on an unchanged process, or to a redesigned core workflow measured in money. The research is consistent that the second choice is where returns appear. Leaders who make that shift now are the ones who will be in the minority that can answer the board's question with a number.

Building your 2027 AI budget? DaCodes can help your team identify the two or three core-function opportunities with the clearest financial case and design them for production. → Talk to DaCodes

Sources 

• BCG, AI is starting to pay off: Applied AI Index 2026, Sep 30, 2026 (press release; full report not opened).
• BCG, The Widening AI Value Gap: Build for the Future 2025, Sep 30, 2025.
• McKinsey, The state of AI in 2026: On the road to ROI, Aug 25, 2026.
• PwC, 29th Global CEO Survey, Jan 19, 2026.
• Deloitte AI Institute, The State of AI in the Enterprise, 2026 (publication date not shown on page).
• Wharton / GBK, Accountable Acceleration: Gen AI Fast-Tracks Into the Enterprise, Oct 28, 2025.
• Stanford HAI, AI Index 2026, Economy chapter.