About 80% of companies are using GenAI. And 80% of them report zero bottom-line impact. “GenAI paradox”, anyone?
Everyone’s talking about AI transformation, but the P&L statements tell a different story entirely.
McKinsey’s latest report “Seizing the agentic AI advantage” on agentic AI unearths some uncomfortable truths about where most organizations stand today. And it’s not pretty.
Below, I’ve distilled 4 critical insights from their analysis that should make every CEO think twice before jumping headlong into AI.
Let’s dive in.
1. The Great GenAI Illusion
Nearly eight in ten companies report using gen AI. Yet, just as many report no significant bottom-line impact.
Most organizations have gone for the easy wins like enterprise-wide copilots that make individual employees slightly more productive. Think of it as the AI equivalent of giving everyone a faster stapler. Sure, documents get processed quicker, but the fundamental work hasn’t changed.
Meanwhile, the vertical, function-specific applications that could actually move the needle, remain stuck in pilot purgatory. Fewer than 10 percent of use cases deployed ever make it past the pilot stage.
Why? Because horizontal solutions are plug-and-play. Vertical ones require companies to actually rethink how they work. And that’s hard.
2. AI Agents Are the Game-Changer
AI agents offer a way out of the gen AI paradox. Agents can automate complex business processes, thereby turning gen AI from a reactive tool to a proactive, goal-driven virtual collaborator.
Unlike the chatbots we’re used to, agents don’t just respond when prompted. They plan, remember, adapt, and execute, all without human intervention.
McKinsey shares a case study: A bank used AI agents to modernize 400 pieces of legacy software (originally budgeted at $600+ million). The impact? More than 50% reduction in time and effort.
Another example: A research firm replaced 500+ manual data quality workers with agents that could identify anomalies and generate insights automatically. Potential savings: $3 million annually with 60% productivity gains.
This is structural disruption that AI promises. And, this is what firms must aim for.
3. Success Demands Complete Process Surgery, Not Band-Aid Solutions
Unlocking agentic AI’s potential requires more than plugging agents into existing workflows. It calls for reimagining those workflows from the ground up—with agents at the core.
Think of it this way. You can use a Ferrari to drive to the grocery store on city streets, but you’re not really using what it’s meant for. Cramming agents into legacy processes delivers only modest improvements.
By redesigning the entire process around agent capabilities, parallel execution, real-time adaptation, elastic scaling, you’re looking at significantly higher efficiency gains.
The McKinsey example that drives this home: A call center using AI for assistance sees 20-40% improvements. But reimagine the entire process where agents proactively detect issues, auto-resolve 80% of incidents, and only escalate exceptions to humans? That’s the transformation potential here.
Most organizations aren’t ready for this level of surgery. They want AI to make their current mess work better, not fundamentally challenge how they operate.
4. This Is a CEO-Only Decision
The moment has come to bring the gen AI experimentation chapter to a close. And this is a pivot only the CEO can make.
Why? Because successful agentic AI transformation requires dismantling organizational silos, redesigning core business processes, and fundamentally shifting how humans and machines collaborate. You can’t pilot your way to that kind of change.
Fewer than 30 percent of companies report that their CEOs sponsor their AI agenda directly. That has to change.
The companies getting ahead aren’t just deploying agents, they’re rewiring their entire organizations around them. Moderna, for instance, merged its HR and IT leadership, recognizing that AI isn’t just a tech tool but a workforce-shaping force.
As McKinsey puts it: This is a moment of strategic divergence. Organizations that treat this as another IT project will find themselves competing against fundamentally different operating models.
What This Means for You
Unlike the gentle productivity boosts we’ve seen with ChatGPT and copilots, AI agents are about to reshape entire business processes.
The organizations that figure this out first will redefine their industries, not just gain efficiency.
As for the rest? They’ll be competing against companies that operate at radically different speeds and scales.
Musings by Madhav 
Leave a comment