Enterprise architecture’s value proposition at a turning point – towards architecture practice for the AI era  

A few years ago, when I wrote about the roots of enterprise architecture thinking and where it might be heading, the world looked very different. I could not have imagined that only a few years later, we at Gofore would be rethinking some of the fundamental questions of architecture consulting – this time on an entirely new playing field. Rapid advances in AI are reshaping knowledge work and its economics on an almost historic scale. For architecture professionals, this creates plenty to consider: both new opportunities and unresolved challenges. 

Architecture’s unfulfilled value proposition 

Researchers have spent decades analysing the benefits of enterprise architecture [1], its value creation mechanisms [2] and its success factors [3]. Yet in the everyday reality of organisations, the promise has repeatedly gone unfulfilled. Architecture practice still suffers, at least to some extent, from a gap between expectations and realised benefits [4]. 

The typical reasons why architecture practices fail are familiar to every professional in the field – themes discussed so often by coffee machines and in LinkedIn threads that they have become clichés. The same obstacles recur: descriptions are too laborious to produce and too expensive to keep up to date; communication lacks a shared language, or uses a language that is somehow wrong for one stakeholder or another; and the work remains at the level of theory instead of informing concrete decisions. And if all else fails, there is always the lack of management support – because leaders are not interested in architecture itself, but in the outcomes it enables. 

Importantly, many challenges in architecture practice have resulted either from unsuitable choices of approach or from straightforward cost–benefit problems – not from fundamental flaws in the architectural thinking behind the practice. Systems thinking, which helps us understand wholes and make better decisions based on that understanding, remains a sound foundation for architecture work. What if new technology could make systems thinking more accessible by removing constraints once assumed to be permanent? 

AI clearing obstacles from architecture work 

Discussion about the future of knowledge work is often framed around making tasks more efficient – or even replacing them – with AI technologies. I find the opportunities that emerge when new technology removes the traditional barriers to successful architecture work more important. If the cost of producing, maintaining and using architectural understanding falls significantly, architecture’s value proposition has a real chance of returning – perhaps on a more sustainable foundation than before. 

The mechanisms through which architecture creates value are closely connected to the tradition of knowledge management. It is therefore useful to examine the opportunities created by new technology specifically through the lens of knowledge management processes. Work that once depended heavily on an individual architect’s expertise and available time is becoming more affordable and scalable through AI-based technologies. Natural language processing (NLP) makes it easier to extract architectural understanding from unstructured material. Large language models (LLMs) enable stakeholders to interact with architecture content in their own language. Machine-learning (ML) methods help identify patterns in architecture data, while information management based on shared ontologies supports a coherent view of an organisation’s architecture. Technology promises both to democratise the use of architecture knowledge and to free architects from creating, maintaining and interpreting deliverables so that they can focus on the higher-value work that has long been expected of them. [5] 

We know that the value of architecture is only indirectly derived from its deliverables. Value is created in everyday organisational life by building shared understanding and directing resources towards the right, shared goals [1, 2]. Success requires more than efficient architecture work. It also calls for an organisational culture that is open to architectural thinking, a shared language, and strong communication and leadership skills [3, 4]. AI clearly has the potential to strengthen these conditions, but many questions remain unresolved [5]. 

AI is not a simple answer to decades-old organisational challenges. Architectural interpretation, judgement and creativity – and the ability to turn understanding into organisational action – remain deeply human processes. Responsibility for decisions cannot be outsourced to models built on statistical probabilities. As producing information becomes cheaper, scarcity moves from production to verification. The bottleneck does not disappear; it shifts. As information volumes grow, people face an even harder prioritisation challenge: which signals deserve action, and where should the organisation direct its resources? 

Updating architecture practice for the AI era 

At Gofore, our approach to AI-assisted architecture work has been pragmatic. We have followed developments with great interest while keeping our feet on the ground and recognising both the opportunities and limitations of AI technologies. We have broken architecture practice down into its component parts and identified AI opportunities and use cases – both those where AI already performs well and those where human contribution will remain decisive. Beyond technological capabilities, a key question is what the AI era demands from operational and information-management capabilities when they must serve not only human architects but architecture agents as well. 

One thing is clear: a meaningful step change in architecture practice requires more than automating individual tasks and phases. The entire value stream of architecture work must be dismantled and reassembled as we move from isolated AI-assisted tasks towards the capabilities required by genuinely AI-native – or even agentic – architecture practice. The work is still in progress, but the pace of technological development makes this the most exciting moment for architecture practice in years. What was impossible yesterday may already be possible tomorrow. 

References 

1. Tamm, T., Seddon, P. B., Shanks, G., & Reynolds, P. (2011). How does enterprise architecture add value to organisations? Communications of the Association for Information Systems, 28(1), 141–168. 

2. Niemi, E., & Pekkola, S. (2020). The benefits of enterprise architecture in organizational transformation. Business & Information Systems Engineering, 62(6), 585–597. 

3. Ylimäki, T. (2006). Potential critical success factors for enterprise architecture. Journal of Enterprise Architecture, 2(4), 29–40. 

4. Gong, Y., & Janssen, M. (2023). Why organizations fail in implementing enterprise architecture initiatives? Information Systems Frontiers, 25(4), 1401–1419. 

5. Toumi, A., Fosso Wamba, S., & Hafsi, M. (2025). Enterprise architecture as a knowledge management discipline: evolution, challenges and AI-enabled future. Journal of Enterprise Information Management, 1-26.