The primary function of IT is evolving

An ode to my dear friends in IT.

I've spent the last 15 years working alongside highly technical people, and I want to be clear up front: I'm not one of you. I respect you all too much to pretend otherwise. I'm approaching this from the outside, in the hope that an outside view is actually useful to those navigating what AI is doing to the IT profession right now.

My background is advertising, communications, and marketing. My work lives at the intersection of organizational change and human behavior — how people experience change, resist it, adapt to it, and eventually become it. Over the years I've picked up enough of the technical language to get by, though I'd never claim to understand all of it. What I do understand is patterns, and for the past several years my team and I have had a front-row seat to one of the biggest shifts in enterprise technology in decades, partnering closely with one of the world's largest tech companies as it reshapes how organizations build and run their IT environments.

Lately it's reminded me of something from earlier in my life.

I grew up in New Zealand, where rugby isn't a sport so much as a load-bearing part of the culture. I never played an official game, but I got my bell rung plenty at lunchtime and on the beach with my mates. I remember the specific feeling of catching the ball and looking up to see someone bigger, faster, and considerably more determined charging straight at me. Time compresses in that moment. You're suddenly responsible for what happens next, and you don't yet know what the right call is. Pass too early and you lose control of the play. Hold on too long and you absorb the hit. Hesitate and the opportunity is just gone.

It's not only physical pressure. It's decision pressure.

That feeling has come back to me recently — thankfully not literally. It's close to what I'm seeing across IT organizations right now. Leadership has passed the ball, expectations are enormous, and the environment has changed. Dramatically.

There's a concept in biology called epigenetics: how environmental factors change the way genes are expressed without touching the underlying DNA. Identical twins are the clearest illustration — same genetic blueprint at birth, but separate them into different environments and, over time, they diverge. Different stresses, different development, sometimes different capabilities entirely. The DNA never changed. The environment did, and the environment decided how the genes got expressed.

Organizations work the same way. They have structural DNA — roles, responsibilities, systems, patterns of operation — that isn't fixed so much as adaptive, evolving to match whatever environment it's operating in. For decades, IT's environment was software that was heavy and static, infrastructure that was expensive and slow to change, technology that supported human work without doing the work itself. IT got exceptionally good at operating in that environment: deploy systems, maintain infrastructure, secure access, keep the lights on. It became the architectural backbone of the modern enterprise.

That environment just changed. Not gradually — suddenly. Software isn't static anymore, and intelligence isn't confined to applications. Agents — software entities that can perform tasks, make decisions, and act on systems, sometimes on a human's behalf — are starting to operate alongside people. Capabilities that used to require an entire application can now be spun up in an afternoon. This doesn't invalidate IT's role. It changes how that role has to be expressed. Call it corporate epigenetics: the DNA — architectural responsibility, systems stewardship, governance of access and security — is intact. The environment it operates in isn't the same environment anymore, so the expression of the role has to evolve with it.

Recognizing that is the first step, and it's a bigger one than it sounds. This isn't another tool or platform showing up. It's a structural shift in how capabilities get created, deployed, and operated inside a company. Agents need identity, access, governance, lifecycle management — monitored, updated, secured, integrated, the whole architectural apparatus IT already knows how to build. IT isn't just responsible for the systems employees use anymore. It's becoming responsible for the infrastructure intelligent agents use to get work done — what I'd call agentic infrastructure: the identity frameworks, governance models, orchestration layers, and architectural standards that let agents operate safely inside the enterprise. That infrastructure won't emerge coherently on its own. It has to be designed, and I don't think there's another function in the enterprise better positioned to design it than IT.

The alternative is fragmentation, and IT has seen this movie before — it's shadow IT, except the barrier to entry just dropped again. Agents are lightweight, accessible, fast to stand up, and easy to acquire outside any procurement process. That's not business units being reckless. It's business units being adaptive, feeling the same environmental pressure and reaching for the fastest available answer. But agents interacting with systems without a shared identity framework or governance standard don't just create security risk — they create structural risk that compounds quietly until it isn't quiet anymore. The answer isn't to slow the emergence of agents down. It's architectural stewardship: building the infrastructure that lets them emerge coherently instead of accidentally.

There's a deeper layer to this too. When an organization trains agents on its own data, workflows, and decision patterns, it isn't just automating tasks — it's externalizing institutional knowledge into something that operates. Those agents become part of the company's intellectual fabric: accumulated knowledge and competitive differentiation, expressed as infrastructure rather than tribal memory. Which means IT isn't only operating systems anymore. It's becoming the steward of the infrastructure that contains the organization's intelligence itself. That's a genuine expansion of the role, and I think it's mostly upside if IT chooses to see it that way. The center of gravity is moving: fewer of the decisions that matter most will come through a monolithic vendor deployment, and more of them will live inside the enterprise's own architecture, with IT as the caretaker of that internal ecosystem.

From the outside, all of this can look like disruption. From the inside, it's an expansion. IT's responsibility for deploying and maintaining the systems a company depends on doesn't disappear — it grows to include architecting and governing the agentic infrastructure that intelligent systems need to run safely. No other function has the architectural visibility, the governance mandate, or the systems-level view to design that coherently. This is IT's moment to become the curator of the enterprise's operational future, not because someone assigned it that job, but because nobody else can actually do it.

I think back to those moments holding a rugby ball, watching someone considerably larger charging toward me. What I didn't understand then, and do now, is that the player holding the ball isn't only under pressure. They're in control. In rugby, that's usually the number 10 — the fly-half. Not the biggest or fastest player on the field, but the one reading the game, controlling the tempo, deciding when to pass and when to run and when to make space. The one who shapes the outcome.

The environment around IT has changed, and it's moving fast. The pressure is real. But the ball is in IT's hands now, and with it comes the chance to architect what happens next.

D.