Vinod Bijlani is an AI practice leader at Hewlett Packard Enterprise.
I consistently see organizations asking, “What can we do with AI agents?” That’s the wrong question. It starts with the technology and works backward, and that’s exactly why so many AI initiatives stall after the pilot.
The right question is, “Which processes, if reimagined with AI, would most transform our customer experience and our economics?” That question starts with the business. According to McKinsey’s “State of AI” report, organizations that have moved beyond pilots to scaled AI deployment are seeing gains in productivity and revenue. But scaling only happens with a coherent strategy behind it, and most enterprises don’t have one yet.
From Copilot To OpenClaw: The Agentic AI Moment
At GTC 2026, Nvidia CEO Jensen Huang was unequivocal: “Every company in the world today needs to have an OpenClaw strategy, an agentic systems strategy.”
OpenClaw is to agentic AI what ChatGPT was to generative AI: the moment it stopped being theoretical and became something every organization had to reckon with. Organizations that ignore this shift risk being outpaced (registration required) by agent-native competitors within the next year or so. OpenClaw is the Linux of the agentic world. The infrastructure layer has arrived. The only question now is what you build on top of it.
Five Pillars Of An Enterprise Agentic AI Strategy
Through my work leading AI and cloud transformation initiatives, I’ve come to realize that AI strategy is far more than identifying use cases. It is five interconnected pillars: business strategy, people strategy, data strategy, technology strategy and governance.
1. Business Strategy: Starting With The Operating Model, Not The Technology
Agentic AI doesn’t just automate work; it forces a redesign of how work flows. Sequential handoffs between departments become unnecessary when agents orchestrate cross-functional tasks end-to-end.
McKinsey’s research on AI-era operating models describes an emerging pattern: Humans set strategy and handle exceptions; agents execute and coordinate everything in between. The enterprises getting this right don’t start with a list of AI use cases. They start by asking which customer journeys are broken, which operations are slow and what they would redesign if constraints disappeared. Then they bring in the technology.
2. People Strategy: From Execution To Oversight
This is the pillar most organizations underinvest in. Agentic AI shifts human roles from doing the work to directing and validating it. That is a significant cultural shift, not just a training exercise. Leaders who treat it as a change management problem, not a technology rollout, are the ones who see adoption stick. According to BCG, companies that achieve the most AI value are four times more likely to have structured learning programs in place than their peers, underscoring why intentional upskilling, not organic adoption, drives enterprise-wide AI success.
3. Data Strategy: What Your Agents Actually Run On
Here is something I see constantly: Organizations rush to deploy agents on top of data they haven’t cleaned, connected or governed. The agent looks impressive in a demo, then makes bad calls in production because it is working with fragmented or stale information. Gartner consistently identifies data quality and governance as the top barrier to AI value realization. Fix the data estate first. Interoperability, lineage tracking, access controls and clear ownership are not optional foundations. They differentiate agents that scale from agents that stall.
4. Technology Strategy: Build, Buy And Know The Difference
The build versus buy decision is more nuanced than it used to be. For differentiated capabilities where your proprietary workflows or sensitive data are involved, building may be the right call. For broadly applicable use cases, configuring platforms on top of foundation models is faster and often cheaper.
What I recommend regardless of path: pilot small, validate reasoning accuracy and failure modes before scaling and establish a center of excellence to standardize how agents are designed and evaluated. IBM’s analysis on this shows that testing pipelines, security controls and monitoring cannot be retrofitted after deployment.
5. Governance Strategy: The Pillar Nobody Wants To Talk About Until Something Goes Wrong
Autonomous agents make sequences of consequential decisions without a human in the loop. That is the point. It is also the risk. Enterprises need clear accountability frameworks before deployment: who is responsible when an agent takes a harmful action, what the escalation path looks like and how performance is measured against real business outcomes. The EU AI Act makes traceability and explainability a legal requirement for high-risk systems. Governance is not the brake on innovation. Done well, it is what makes scale possible.
The Evidence Is Already Here
I would say that we are currently in the “experimental phase” of agentic AI. While true end-to-end organizational autonomy is not yet mainstream, the foundational architecture for it is being stress-tested in high-stakes environments.
In financial services, JPMorgan Chase has used AI to automate legal-document analysis and compliance-related reviews.
In healthcare, Epic is incorporating generative AI and agent-like capabilities into workflows such as prior authorization, patient messaging and care coordination, with the goal of reducing administrative burden on clinicians and allowing more time for patient care.
In retail, Walmart has expanded its use of AI across forecasting, inventory replenishment, logistics and supply-chain operations. The company has also begun exploring agentic AI applications, making it one of the largest enterprises experimenting with AI-driven operational decision support.
Why The Window For Strategic Advantage Is Now
Agentic AI is not a feature upgrade to your existing stack. It is a redesign of how work gets done. The organizations I’ve seen succeed treat it as a business transformation program, not an IT initiative. Executive sponsorship, clear metrics, a coherent strategy across all five pillars and the discipline to pilot before scaling. That mindset, more than any model or platform, is the real competitive moat.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
