On AI adoption, organizational design, structural coherence, and the decisions that matter most right now.
Potentially — but doing so introduces risks most organizations are not prepared to manage.
AI can increase efficiency and reduce manual effort, but that does not mean it should be treated primarily as a headcount-reduction tool. AI does not replace the full value people provide: judgment, context, accountability, ethical discernment, and the ability to adapt when reality does not match the model.
In most cases, the stronger approach is augmentation — using AI to increase human capability rather than prematurely removing the humans who make the system resilient. Humans working in concert with AI will usually create more durable value than AI operating alone.
The real danger is structural. Many organizations are still relying on people to quietly compensate for weak workflows, unclear decisions, broken handoffs, and missing governance. If those people are removed too early, AI may not eliminate the problem — it may amplify it.
A better question is not, "How many people can AI replace?" It is, "How do we use AI to increase capability without increasing fragility?"
Because AI investment by itself does not create business value.
Many organizations are seeing local gains — faster output, reduced manual effort, improved individual productivity. But those gains do not automatically translate into better revenue, stronger margins, or meaningful enterprise performance.
The problem is rarely the technology. It is the surrounding system. AI introduced into fragmented workflows, unclear decision structures, and uneven adoption patterns will produce more speed without producing more value. The organization gets busier. The results stay flat.
This is what organizational entropy looks like under AI acceleration. The inefficiencies that existed before — the coordination gaps, the unclear ownership, the decisions that take too long — do not disappear when you add AI. They scale.
The organizations that do see returns from AI are not necessarily the ones with the most sophisticated tools. They are the ones with enough structural coherence that increased speed has somewhere to go. AI amplifies what the organization already is. If the structure is fragmented, AI makes fragmentation faster. If it is coherent, AI compounds that coherence.
Revenue follows coherence. AI just accelerates the distance between them.
This view is grounded in both working experience and a consistent pattern across current research.
McKinsey's 2025 State of AI survey found that only 6% of organizations are achieving meaningful EBIT impact from AI, and nearly two-thirds have not yet begun scaling AI across the enterprise. BCG's Widening AI Value Gap report found that only 5% of companies are achieving AI value at scale — while 60% report minimal revenue or cost gains. IBM's 2025 CEO study found that only 25% of AI initiatives have delivered expected ROI, and only 16% have scaled enterprise-wide.
The counterpoint matters too. Deloitte's research shows that nearly three-quarters of organizations' most advanced AI initiatives are meeting or exceeding ROI expectations — so real returns are achievable. But the same research found that revenue growth from AI remains aspirational for most: 74% of companies hope to grow revenue through AI, while only 20% are already doing so.
The pattern is consistent. Local AI wins are real. Enterprise-wide value is not automatic.
Peak Agility's position is that structural coherence is the determining factor — and the research supports it directly. Deloitte found that companies which redesign their workflows are twice as likely to exceed ROI expectations, and that 93% of AI budgets go to technology while only 7% goes to the behaviors and capabilities needed to use it well. BCG identifies clearly defined outcomes, workflows, and decision structures — not more sophisticated tools — as what separates value leaders from the rest.
The evidence does not challenge the assertion. It substantiates it.
Have a question that isn't answered here? Or want to apply this thinking to your specific situation?
Start a Conversation