Editorial correction · November 3, 2026: An earlier version subtracted 25 from 85 incorrectly. The arithmetic is corrected here; the original report and population comparability still require source verification.
The Federal Reserve Bank of Atlanta published a study in May 2026 that put a number on something most business leaders feel but cannot articulate. The top 10% of firms investing in AI plan to spend at least $2,800 per employee this year. The median firm plans to spend $200 or less. That is a fourteen-to-one ratio.
If you run a business and your AI budget falls closer to the $200 end, the instinct is to spend more. Before writing a bigger check, look at what the top spenders have been building underneath those numbers for the past 18 months while everyone else was running pilots. Their budgets reflect an operational advantage that also needs closing.
Compare spending with returns
Gartner projects worldwide AI spending will hit $2.59 trillion in 2026, up 47% from last year. That is an industry-wide commitment at a scale most people are still trying to absorb.
BCG estimates that roughly 5% of companies are generating outsized value from their AI investments. Close to 60% report little material return despite real spending. The gap between capex and revenue growth is running at about 46%, according to Allianz Research. For most companies, returns remain well below the money going in.
AI spending needs operational change behind it to produce value for your company as well as revenue for your vendor.
The Adoption Gap Is a Practice Gap
IBM’s 2026 CEO study found that 85% of employees now have access to AI tools. Only 25% use them regularly. Subtracting those reported percentages gives 60 percentage points, provided they refer to the same population.
Publicis Sapient surveyed 1,550 AI decision-makers and found that 73% say AI is used regularly across business processes. But only 10% say AI is core to how their business actually operates. Turning regular use into operational integration is where the value lies.
Writer’s enterprise report puts it more bluntly: 79% of organizations face challenges adopting AI, a double-digit increase from 2025. And 84% of companies have not redesigned jobs or workflows around AI. Their purchased tools still sit on top of the old work.
Why This Compounds
The organizations at the top of that 14x spending curve are building operational muscle through repeated practice alongside their investment.
Their teams have developed habits around AI. Their workflows incorporate agents and automation at the process level. Their managers know how to direct AI work as well as approve AI purchases. That institutional knowledge compounds every quarter. Each new tool, each new model release, each capability improvement lands on top of an operational foundation that is already working.
For the organizations at the bottom of the curve, each new release lands on the same broken foundation. Workflows, habits, and ownership of process change still need to be established before teams can use the available capabilities well.
This is why the practice gap compounds: the practiced side gets faster while the unpracticed side stays still.
The 54% reporting organizational friction
Writer’s report included a statistic that deserves more attention: 54% of C-suite executives say that adopting AI is tearing their company apart.
That points to a change management failure happening in real time. Organizations tried to layer AI on top of existing operations without changing those operations. The result is friction, confusion, and internal resistance at the executive level.
The companies in that 5% generating outsized value addressed this friction by redesigning the work before deploying the tool. They asked the operations question first. What changes in how we work? Who owns the new workflow? What do we stop doing? The tool came after the answers.
What This Means for the Next Six Months
If your organization is in the $200-per-employee range and you are planning to increase the budget, connect the additional licenses to a plan for building operational capability.
The 14x spending leaders built their advantage through three operational decisions. First, someone with authority over how the work gets done owns the AI workflow in every department. Second, they redesigned at least one core process around AI before scaling, putting a production workflow in place of the old process. Third, they measured how the output changed at the operation level, going beyond tool login counts.
Teams at budgets below $2,800 per employee can make all three decisions. Most organizations still need to make them.
The gap is fourteen to one today. In six months, organizations that continue with unchanged workflows will face leaders with six more months of practice. Building that experience takes time, so begin before Q4.