Enterprise Architecture Is Dead
By Gareth Humphreys
Like a lot of people who started their tech career in the mid-90s, I ended up in Architecture by accident.
Maybe it was because I wasn’t operating at the right level of organisational or technical complexity, but I’d never actually met a Solution or Technical Architect until I became one myself.
I’d love to say it was the result of a promotion or a long-trodden path that culminated in elevation to some mythical architecture guild, but the truth is simpler: job boards had started advertising this new-fangled “Architect” role, and it came with a pay bump. That was all the incentive I needed.
I landed my first architect role straight from a hands-on technical gig. “Architect” certainly sounded more impressive than “Technical Manager,” even if I felt like a bit of a fraud - especially compared to those poor souls in construction who spent seven years studying to design real things in the real world.
But how hard could it be?
I was already churning out design documents. Conceptual, Logical, Physical. Traceability matrices, entity relationships, governance forums, cost models. The works. Then I discovered TOGAF and Zachman. And that’s when I saw the next level - the frameworks, the layers of abstraction, the comfort of structure. A known-good approach to shaping systems around business needs. Elegant. Logical. Brilliant, when it was allowed to breathe.
But here’s the rub. Then - and especially now - it’s rarely given the oxygen it needs.
Instead, architecture turns into performance art. Diagrams for the sake of diagrams. Governance theatre. A sea of chin-stroking and “safeguarding the future state” while the business gets twitchy and competitors ship products.
I’ll say this clearly: I love architecture. I love the discipline, the complexity, the puzzle-solving. How do you get all these moving parts to work together across functions, tech stacks and budgets, without dropping the ball?
But I’ve also seen the dark side. Vanity projects masquerading as strategy. Over-engineered frameworks nobody reads. Programmes like the EU’s €1 billion Enterprise Architecture Modernisation Programme (EAMP), which (depending who you ask) spent eight years generating heat but not much light, before being quietly culled in 2021.
The Harsh Reality
EAMP wasn’t a one-off. It was a symptom. It was a case study in how traditional architecture models struggle when reality moves faster than the framework. But why?
1. Rigidity in a Fluid World
Enterprise Architecture came of age in a slower world. Annual planning cycles. Predictable tech stacks. Stable teams.
That world’s gone.
Now, markets pivot overnight. Boards want answers this quarter, not next year. And AI is only accelerating that pressure.
Try telling an exec frothing over GenAI’s potential that you need six months to define some “guardrails” before experimenting.
2. Complexity That Kills Innovation
We were told that frameworks like TOGAF brought clarity. In practice, many brought bureaucracy.
The more complex the model, the harder it is to move at speed. And when you bolt AI into the mix - a tech that thrives on iteration, experimentation and short cycles - that complexity becomes a straightjacket.
You don’t need a committee to validate a chatbot POC. You need clear governance that doesn’t kill momentum.
3. The Tech-Business Disconnect
Architecture should be the bridge between business intent and technical delivery. Too often, it becomes a vacuum.
Generative AI is forcing organisations to think differently, about content, decision-making, even compliance. But if your architecture function is still stuck on abstract capability maps while the business is piloting LLMs in production, you’re already behind.
4. Innovation at a Snail’s Pace
AI moves fast. Architecture doesn’t. That’s the tension.
When change control takes weeks, and design sign-offs take longer than the tech itself takes to become obsolete, you’re not governing, you’re obstructing.
5. Digital Transformation’s Worst Enemy
We used to think Architecture was the enabler of transformation. In many cases, it’s become the bottleneck. Especially when it insists on working like it’s still 2005…when today’s technology stacks are composable, API-first, cloud-native and increasingly AI-powered.
Transformation is fluid. Architecture, too often, is not.
Enter AI – And the Fork in the Road
Here’s where it gets interesting.
AI doesn’t kill architecture, but it does change the job.
Because the tools we now have - AI copilots, intelligent agents, LLMs that can generate code, docs, models - are already starting to reshape what architects do and how they do it.
Good architects are already:
Using AI to accelerate design artefact generation (not to avoid design, but to scale it).
Using AI to simulate trade-offs and generate options, not just draw static boxes.
Using AI as an interface between human intent and system capability, especially where non-technical stakeholders are involved.
Bad architects are still drawing 14-layer abstractions that say nothing, to no-one, about what to do next.
So What Now?
The answer isn’t to get rid of architecture. It’s to evolve it. Quickly.
Modern architecture needs to be:
Outcome-led, not artefact-driven
Enabling, not controlling
Fast, not perfect
Deeply integrated with AI-first thinking - not sitting outside of it
The job of an architect now is to keep things coherent while letting innovation breathe. To avoid chaos without enforcing stagnation. To make the AI opportunity safe without making it slow.
It’s not easy. But it’s necessary.
Final Thought
The death of traditional EA isn’t the death of architecture. It’s the rebirth of something better that’s leaner, faster, smarter.
And AI isn’t a threat to this new kind of architect. It’s their most powerful tool. Unless, of course, they cling to the past. In which case… well, we’ve all seen how that story ends.