Enterprise Systems · AI · Operational Transformation
David
Gorman
The models are new. The operational constraints are not. Writing on enterprise AI, infrastructure, and navigating the organizational realities that help technology transitions actually work.
"The interesting work is usually at the point between what technology makes possible and what organizations are actually ready to adopt. That's where I've spent my career."
About
How the perspective formed
Enterprise technology has moved through several eras -- networking infrastructure, storage and data protection, cloud platforms, distributed data systems, and now generative and agentic AI. Each one exposed the same underlying reality: the technology moves faster than the organizational and operational systems built to absorb it. Enterprise technology doesn't fail because capability is absent. It fails because organizations struggle to absorb and operationalize the change surrounding it.
The recurring problem across all of them: architecture carries operational consequences that organizations don't recognize until they're already committed. Latency compounds. Governance gaps surface at runtime. Data that seemed positioned well enough turns out not to be. I've worked at that boundary -- product management, product marketing, portfolio strategy, and GTM at AWS, Pure Storage, Brocade, EMC, and Juniper -- helping organizations understand what they're trying to achieve and what the infrastructure decisions they're making will actually mean in operation. I carry this experience into positioning and messaging that helps organizations understand the operational consequences of the decisions they're making and connect technology choices to real business outcomes.
Generative and agentic AI are genuinely different from what came before -- the reasoning dynamics, the context dependencies, the inference economics. But the organizational constraints underneath aren't new: data that isn't positioned correctly, governance that doesn't extend to runtime, trust that hasn't been operationally established. Most of the writing here is an attempt to make those constraints visible before organizations learn them the expensive way.
Writing
Articles & Essays
On enterprise AI, infrastructure, why technology transitions are harder to operationalize than they appear, and how to work through them.
Recurring Observations
Recurring ideas worth naming
Certain patterns keep surfacing across technology transitions, organizations, and scales. These are attempts to name them precisely enough to be useful.
AI-First Discoverability
Generative AI now sits between content and audience. Whether information gets surfaced depends on how it's structured and whether it aligns with intent -- not just whether it exists.
Context Engineering
Context isn't a prompt feature -- it's an operational layer that has to be designed, governed, and maintained. What reaches an AI system at inference time determines the reliability of everything it produces.
Architecture Determines Outcomes
Infrastructure decisions made early constrain what's operationally possible later. Latency, data proximity, and retrieval design aren't implementation details -- they're the conditions under which AI systems either hold up in production or don't.
Content vs. Assets
Content is the strategic idea -- the positioning, the argument, the framework. Assets are what gets derived from it. Conflating the two produces fragmented messaging. Separating them is what makes positioning hold across channels and scale.
Operational Trust
AI systems don't just require secure access to data. They require confidence in how data is retrieved, assembled, governed, and acted on. As AI becomes operational infrastructure, trust shifts from static access control toward continuous visibility and organizational confidence in what's running on their behalf.
Store Once, Use Many
Siloed, duplicated, or access-fragmented data creates compounding friction across every AI workload that depends on it. Where and how data is stored is a decision about how much operational overhead future AI systems will carry.
Speaking & Media
Conversations
Interested in conversations about enterprise AI systems, operational strategy, and why technology adoption is harder than the capability itself suggests. Occasionally available for podcasts, panels, and industry discussions.
Enterprise AI Operationalization
Most enterprise AI pilots don't fail because the model is wrong. They stall because the operational, data, and organizational conditions required to run AI in production were never established. What that gap looks like in practice, and what closing it actually requires.
Inference Economics at Scale
Latency, cost, data proximity, and retrieval design are not implementation details -- they are the variables that determine whether an AI system can actually run at enterprise scale and hold up under production conditions.
AI-First Discoverability
Generative AI systems have become the first layer of discovery for enterprise content. The structure and intent-alignment of information now determines what gets surfaced -- and most content strategies were built for a different model of how information gets found.
Experience
An Operational Systems Arc
From network infrastructure through storage, cloud, and data systems into generative AI -- each transition exposed the same organizational friction: the operational systems needed to absorb a new technology aren't in place when the capability arrives. The work across these roles has been helping organizations close that gap.
Contact
Interested in conversations where enterprise AI systems, operational strategy, and organizational realities intersect.
If you're working on problems at the intersection of enterprise AI, infrastructure strategy, and operational transformation -- I'm interested in the conversation.
Connect on LinkedInWriting and working at the intersection of enterprise AI, operational systems, and organizational transformation.