AI is already drafting your diagram
AI can now sketch a capability map, summarise a 60-page strategy paper and flag gaps in a roadmap before your coffee goes cold. So where does that leave the enterprise architect? Not out of a job, but definitely in a different one. The architects who thrive next won’t be the ones who model fastest. They’ll be the ones who think widest and go deep where it counts. In other words, they’ll be T-shaped.
What T-shaped means for EA
A T-shaped architect has two things going on.
The horizontal bar is breadth. It covers business strategy, data, security, cloud, integration and, increasingly, AI literacy. You don’t need to be an expert in all of it. You need enough to ask good questions and spot where a decision in one area will cause pain in another.
The vertical bar is depth. It’s the one or two areas where you’re the person others call, whether that’s data architecture, security, integration or a specific industry.
Breadth earns you a seat at the table. Depth earns you trust once you’re there. Without breadth, you’re a specialist who can’t see the bigger picture. Without depth, you’re a generalist who can’t back up the advice. EA needs both, because the job sits right between the boardroom and the build team.
The skills that gain value
As AI takes on more of the output, the human skills become the main event.
- Systems thinking: Seeing how a change in one place ripples through people, process, data and technology.
- Influence: Getting executives, delivery teams and vendors to row in the same direction, often without direct authority.
- Business-to-tech translation: Turning a strategy into an architecture, and an architecture back into business outcomes.
- Governance: Setting guardrails that keep teams moving instead of slowing them down.
- Judgement: Weighing trade-offs when there’s no clean answer. AI can list the options, but someone still has to own the call.
None of these show up on a tool’s feature list. They’re built through experience, curiosity and plenty of conversations.
The skills AI will absorb
Some of the work is heading for automation, and that’s mostly good news.
Expect AI to handle first-draft models, routine documentation, standards checks and repetitive impact analysis. Tasks that once ate whole afternoons will take minutes.
The catch is that the architect’s role shifts from producing artefacts to validating and steering them. AI can sound confident and still be wrong about your organisation’s context, constraints or politics. Someone has to check the output against reality, which takes enough depth to spot a dodgy model and enough breadth to see what it missed.
Time saved on busywork is time freed for the conversations that actually move things forward.
Why tooling still matters
Here’s the risk with AI-assisted architecture: it’s fast, but fast can get messy. Without a shared, governed model, you end up with AI-generated diagrams scattered across slides and drives, each one slightly different.
That’s where a strong modelling platform earns its keep. At Mitrais, we work with Sparx Enterprise Architect, a multi-user modelling tool that supports UML, SysML, BPMN and other open standards. A few things make it a solid backbone for a T-shaped team:
- Complete traceability from requirements through analysis, design, implementation and deployment, so changes can be tracked and reviewed.
- One shared view of the enterprise. Its high-performance repository handles very large models and lets big teams work from the same picture.
- Built for distributed teams, with integrated version control and cloud-based servers.
- Broad language support, with code generation and reverse engineering across many popular languages.
Think of it this way: AI speeds up the drafting, and the model keeps everyone honest. Breadth and depth both plug into the same source of truth.
Building a T-shaped team
You can’t hire your way to a perfect team, but you can grow one on purpose. Start here:
- Map the skills you have. Plot each architect’s breadth and depth. The gaps, including the single points of failure, show up fast.
- Rotate people across domains. A few weeks with the security or data team does more for breadth than any slide deck.
- Pair specialists with generalists. Depth rubs off when people tackle real problems together.
- Make AI literacy a baseline. Everyone should know what AI does well, where it falls over and how to challenge its output.
- Invest in tool fluency. Training on the platform your team models in, like Sparx Enterprise Architect, means people spend their energy on thinking rather than fighting the software.
- Reward influence, not just output. Recognise the architect who got a hard decision across the line, not only the one who drew the neatest diagram.
Small, consistent steps beat a big restructure.
Where to from here
The EA role isn’t shrinking. It’s moving up the value chain, from drawing the picture to shaping the decision. T-shaped architects, backed by a shared model and sensible use of AI, are well placed for that shift.
If you’d like to talk through your team’s skills or see how Sparx Enterprise Architect could fit, book a free consultation with our Australian-based specialist. Prefer to upskill first? Explore our Sparx System Enterprise Architect Training.
