HBR Just Dropped a Report That Confirms What I’ve Been Seeing in the Trenches
84% of executives believe agentic AI will transform their business. Only 5% have defined success metrics for it.
Harvard Business Review Analytic Services just published their research on agentic AI—surveying 623 executives about expectations, readiness, and results. And honestly?
The findings mirror almost exactly what I see every day. We’ve deployed agentic AI across thousands of customers handling workflows for millions of users.
The gap between belief and execution is staggering.
But here’s what the report gets right that most commentary misses: the organizations actually seeing results aren’t the ones with the best technology. They’re the ones who understood this is a business transformation, not a tech project.
The Readiness Reality Check
The numbers are sobering:
→ 74% say AI is very important to their organization
→ Only 26% say they’re actually effective at using it to achieve business outcomes
That 48-point gap? That’s the chasm between PowerPoint strategy and operational reality.
And it gets worse when you look at workforce readiness. Just 5% of organizations say their people are “very prepared” for agentic AI. Among laggards? 60% say their workforce is “not at all prepared.”
I quoted in the report about something we’ve learned the hard way: “Agentic AI is a forcing mechanism to make sure that we’re well on the journey toward consistently good data.”
Everyone has messy data. That’s not the issue. The issue is treating data cleanup as a prerequisite instead of a parallel workstream.
You don’t have to embark on a multi-year project to get your data right before you adopt agentic AI. The key is to align on which data you need for the specific workflow or job being done. Understand where that data is and ensure its quality and consistency.
Just start there.
What Separates Leaders from Laggards
The report segments organizations into three groups: Leaders (34%), Followers (40%), and Laggards (26%).
The differentiators aren’t what you’d expect.
Leaders don’t have better technology budgets. They have better clarity. Specifically:
Strategy integration: 13% of leaders have agentic AI fully integrated into organizational strategy vs. just 1% of laggards.
Success metrics: 10% of leaders have well-defined success metrics vs. 1% of laggards.
Workforce understanding: 66% of leaders say employees understand the value of agentic AI vs. 24% of laggards.
Notice what’s missing from that list? Technology infrastructure gaps. Data architecture issues. The biggest barriers cited are lack of talent/skills (48%) and no clear roadmap (46%).
The technology works. The human systems often don’t.
The Transformation Trap
This is the part that hits closest to home.
As I told HBR: “This is not a technology project but a business transformation. What you don’t want to do is take an existing business process and put it on a shiny new toy of a tech stack. Then all you’ve done is made an expensive business process.”
I see this constantly. Companies automate broken workflows and wonder why they’re not getting 10x results.
I use what I call “directed autonomy”—where agents can act independently within defined parameters, but humans remain in the loop for oversight, exception handling, and learning.
→ For routine workflows: fully autonomous
→ For context-dependent workflows: shared control
→ For high-impact workflows: always a human in the loop with escalation pathways
We would never let agentic AI create a medical diagnosis on its own. Our aim is to use it to help the clinician make the correct diagnosis—by providing them with all the relevant patient information they need to do so.
The goal isn’t to remove humans. It’s to give them superpowers.
The Results Actually Being Achieved
Here’s what organizations already piloting or using agentic AI are reporting:
→ 36% achieved greater organizational productivity
→ 35% achieved better data-driven decision making
→ 33% achieved cost savings
→ 26% achieved better customer experience
These align with the top benefits organizations are seeking. Which tells me something important: when agentic AI is implemented thoughtfully, it actually delivers.
The disconnect isn’t between expectations and technology capability. It’s between expectations and organizational readiness.
What I’d Add to the Report
A few things that resonated from my own experience:
Data stewardship beats data projects. We’ve adopted a model where data is a trusted asset and we’re all stewards of it. Whether you work in product development, engineering, operations, IT, or on the business side—everyone owns the data from their perspective. We’re all jointly responsible for ensuring it’s relevant, accurate, and has the contextual integrity of what we’re trying to accomplish.
Time to value is the killer metric. The reduction of time spent from ideation to proof of concept, creation, and deployment—to actually generating value—is by far the outstanding benefit of agentic AI. What used to take weeks can be done in a few days. Every part of that value chain is shortened.
Change management matters more than technical deployment. Just because senior leadership wants to capitalize on automation and productivity efficiencies, it doesn’t mean the rest of your organization is along for the ride. There will always be skeptics. Turn those skeptics into cheerleaders and champions. That carries a lot more weight than any internal memo.
The Bottom Line
90% of executives expect most organizations in their industry will use agentic AI in the future.
72% expect it will deliver ROI.
84% believe it will transform their business.
And yet the report makes clear: most organizations are still in exploration mode.
Only 34% are actually using agentic AI in any meaningful way.
The window is still open. But it’s closing faster than most realize.
The winners won’t be the ones with the most sophisticated technology. They’ll be the ones who understood earliest that agentic AI requires reimagining processes, not just automating them.
As John Santaferraro put it in the report: “The winners in every industry will be those that innovate, that don’t just do old things better but that use technology in new ways.”
We’re not just optimizing the old playbook. We’re writing a new one.
What’s your organization’s readiness score? I’d genuinely love to hear—what’s the biggest barrier you’re facing with agentic AI adoption? Hit reply or drop a comment.