Enterprise Architecture is a big word that hangs heavy in the air. What does it actually mean for your organisation, and how does AI fit into the picture? In this talk, which I gave at the HSLU (Lucerne University of Applied Sciences and Arts) in association with the Digital Veterans Association, I explore how Enterprise Architecture, platform engineering, and AI come together as a strategic lever for modern organisations.
What is Enterprise Architecture?#
Let me start with what it is not: it has nothing to do with the Starship Enterprise or Star Wars. Enterprise Architecture is about taking a holistic view of an organisation. Gartner defines it as a discipline for proactively and holistically leading enterprise responses to disruptive forces by identifying and analysing the execution of change toward desired business outcomes.
I visualise Enterprise Architecture as a layered model. At the top sits the Enterprise Vision, which translates into Business Strategy, then Business Architecture. The Business Architecture connects to the Application Architecture through business services. Below that lies the Technical Architecture, supported by methods, skills, tooling, processes, and organisational culture.
Take a Swiss bank as an example. The bank has a vision and business strategy. They offer business services through an e-banking system. Behind that sits the application architecture (the e-banking software), and below that the technical architecture (the core banking system and backends). This layered view across the entire enterprise is what Enterprise Architecture provides.
Gregor Hohpe puts it beautifully and concisely:
“Enterprise Architecture is the glue between Business and IT Architecture.”
The Enterprise Architect: from ivory tower to factory floor#
The stereotype of an Enterprise Architect sitting in an ivory tower, drawing pretty diagrams that nobody cares about, is unfortunately still common. But that is not how it should work.
An effective Enterprise Architect rides the architect elevator, a concept from Gregor Hohpe. You go up to the C-level to discuss business strategy, vision, and IT strategy. Then you go all the way down to the factory floor to work with the people who actually build things. This constant movement between strategic vision and operational reality is what makes Enterprise Architecture effective.
The core responsibilities are: understand the business strategy, translate it into IT strategy together with the CIO, create the “As-Is” picture (what capabilities and applications exist today), define the “To-Be” picture (what the future should look like), and build a roadmap to get there.
Enterprise Architecture is a continuous process, not a project#
Here is something critical: more than 66% of Enterprise Architecture initiatives fail. One likely reason is the word “initiative.” It sounds like a project with a start and an end. But Enterprise Architecture never ends. It is a continuous cycle.
You understand the vision and goals, extract the business strategy, derive the IT strategy with the CIO, create the As-Is picture, define the To-Be picture, build the roadmap, and then govern, refine, and coach. But then something happens. A global conflict, a pandemic, a Crowdstrike incident. The business strategy adapts, which means the IT strategy must follow, and the roadmaps shift.
This is not a project. This is a continuous process that never stops. When done right, Enterprise Architecture reduces cost, accelerates delivery, reduces risk, and makes the organisation more agile.
The digital factory: from strategy to execution#
To show how Enterprise Architecture works in practice, I use the example of a drone company. The board has a vision: expand from small drones to a range of products for higher market share. They maintain a portfolio backlog of prioritised ideas. One idea, drones that carry high loads, gets passed to product management. Product managers break it into features: bigger battery, better motors, software updates. Teams start working.
The critical element: a platform team provides standardised development and delivery environments. New teams can start within minutes instead of spending weeks on setup. After delivery, telemetry data flows back to the teams for continuous improvement and back to the board for strategic decisions about investment.
This is the Digital Factory: lean portfolio management connecting strategy with execution, product management organising work, product teams doing DevOps, and the platform team providing the foundation.
Platforms: where Enterprise Architecture becomes real#
Here is where it gets exciting for Enterprise Architects. The CNCF Platform White Paper describes how platforms provide capabilities: compute, networking, data, identity management, workloads, storage, and more. On top of these capabilities sits an interface with documentation, project templates, golden path templates, a graphical interface, APIs, and CLI.
The key insight for Enterprise Architects: with such a platform, you can codify your guidelines, rules, and policies directly into the infrastructure. Developers do not need to read your governance documents because the governance is already built into the platform. When a developer requests a database, it gets created with the correct naming scheme, the right sizing, proper backups, logging, and monitoring, all with one click.
“With such a platform, you can codify all of your policies, all of your guidelines and rules. Nobody needs to read them because they are already codified and ready to be used.”
This is why platforms are so strategically important for Enterprise Architecture. Gartner and other consulting firms predict that by 2027, all large companies will create their own platforms.
AI in the Enterprise Architecture picture#
Now, where does AI fit? AI is a capability of the platform. It sits in one box of the architecture, but when you zoom in, that box is not small at all.
It contains:
- Platform interfaces: Developer portal, chatbots, CLI, APIs
- Applications: Enterprise chatbots, synthetic data tools, AI coding assistants, knowledge management, productivity assistants
- Tools: Prompt engineering, vector databases, RAG solutions, model lifecycle management (MLOps)
- Model hub: Model registry with versioned models, both self-trained and large language models from providers
- Infrastructure: Compute, storage, and interfaces to AWS, Azure, Google Cloud, and OpenAI APIs
I demonstrated this with the Zühlke Platform Plane, built together with LGT. On the platform, we integrated AI use cases that our developers love: a documentation assistant (a RAG use case over the platform docs), Docker image analysis using a specialised Red Hat LLM, and log analysis using an LLM optimised for that purpose. Each of these was built quickly because when you have a platform with all capabilities in place, creating AI use cases is like playing Lego.
Gartner predicts that by 2027, 40% of platforms will have AI integrated into their software delivery lifecycle. We are already there.
Key takeaways#
- Enterprise Architecture is the glue between business and IT: It ensures the IT strategy is aligned with the business strategy.
- It is a continuous process, not a project: Enterprise Architecture never stops. The business environment keeps changing, and the architecture must adapt.
- The architect elevator is essential: Enterprise Architects must move between the boardroom and the factory floor. Ivory tower architecture fails.
- Platforms codify governance: With a platform, your guidelines, rules, and policies become part of the infrastructure. This is far more effective than documents nobody reads.
- AI is a capability of the platform: It is important, but it is still one capability among many. Govern it and standardise it through the platform.
- Self-service is key: No ticketing systems. Enable teams with a one-stop shop where they have everything they need.
- By combining Enterprise Architecture, platforms, and AI, you create a strategic lever: This boosts performance, enables innovation, and positions your organisation to continuously deliver value.