How AI Will Disrupt the Way Enterprise Architecture Analysis is Done

Architecture analysis is one of the two key activities architects perform within enterprises. This analysis is important in assisting executives in solving their most pressing business problems. However, with the advent of AI tools like Bard, ChatGPT, and others, it is clear that the way architecture analysis is done today is not how it will be done three to five years from now.  

With AI-driven capabilities rapidly advancing and disrupting all sorts of professions, it is no surprise that enterprise architecture will be impacted by this technology trend and forced to evolve in the coming years. But how will AI disrupt the architecture practice? Will it replace architects, allowing executives and stakeholders to get architecture insights instantly rather than from their architecture teams or will this be a technology that enhances the architect toolkit and makes their role even more important? What is clear is that AI will be a large disruptor and change won’t be optional for this profession.

Architecture Analysis Today

To understand how AI is likely to impact architects, we need to first understand the architecture analysis process today.  There are a series of steps, the interactions and activities architects go through to perform the necessary architecture analysis to solve a business problem today. 

It all starts with a stakeholder or sponsor engaging the architect with a business (or technical) problem to analyze. Here are the typical steps that would result from that request:

  1. Meeting with sponsors, stakeholders, and leaders to better understand the business problem or the question they are being asked to answer through architecture analysis.
  2. Interviewing key stakeholders to understand the different pieces of the puzzle. 
  3. Collecting data from various sources, including systems, spreadsheets, and data collection (discovery) activities.
  4. Building a metamodel or a data model from the data they’ve gathered. These multi-faceted models connect all the different concepts, especially when the business problem deals with business applications, processes, and risks.  
  5. Mapping and modeling the relationships that exist between the data itself, the actual applications to their capabilities, and risks to business processes or technology dependencies. When mapping, architects often need to extend their meta-model because as new facets are discovered, new iterations of a model are needed to map out the interconnectivity and interdependencies thoroughly.
  6. Building reporting dashboards. Once architects map and assemble all the data aggregated together through the required mapping and modeling, they build dashboards to report the insights. These insights are incorporated into PowerPoint presentations, Confluence pages, and other content platforms for the stakeholders to digest the information. 

As you can see, today this process is very resource-intensive and can take weeks from the time the architect was engaged to the point where they are able to provide actionable insights.  Because of the length of time elapsed, it is also common for either the environment or the underlying business problem to evolve during the analysis, resulting in multiple iterations of this process..

Architecture Analysis of Tomorrow

The Architecture Analysis process will change dramatically from how it is done today as a result of the AI capabilities that are now becoming mainstream. The processes will become faster, less dependent on the skills/knowledge of architects, and become more technology-driven, with data architecture and AI advancements playing a pivotal role. 

Soon an enterprise’s numerous data sources (CRMs, ERPs, HR systems, finance systems, ITSM systems, etc.) will be accessed through an integration platform or data warehouse, allowing for easier data querying and access from a single location instead of aggregated independently for each analysis request.

Data models won’t need to be designed as much as they will be derived from the operational data itself.. Machine learning and other technologies will examine the affinities and the correlations among data sets and identify where relationships exist, even if it’s not explicitly defined within the data. These technologies can make assumptions about different types of relationships and dynamically map various data sources together in a much more efficient way than manual analysis by an architect at a much larger scale..

The modeling tools will be able to generate architecture diagrams automatically instead of diagrams being manually drawn. The architecture modeling tool will connect into integration platforms and derive the architecture metamodel to show what connective tissue exists within the organization. Those diagrams will be dynamically generated by the tools instead of manually created by architects.

AI tools like ChatGPT, Bard, or the next generation will also provide a new interface where executives and decision-makers can ask questions and get responses with text-based answers or through dashboards. Not only will their responses be rendered dynamically by AI or through BI dashboarding tools, but also the follow-up questions that, in the past, may have required additional architectural analysis. With advancements in AI, decision-makers can obtain faster insights in seconds instead of weeks and be empowered to do their iterative analysis through these tools without taking up excessive resources.  

AI is Coming – How Architects Can Prepare Today

The disruption of AI is coming quickly to the architecture profession. A few months ago, a safe assumption was that this disruption was still a few years away, but recent market developments are accelerating the timelines to within the next 2 years.  What does this mean for architects today? What can architects do now to prepare for changes that are coming? 

The first thing architects need to do is embrace the change. They should not fight it or assume they can continue to do their job in the same old ways. The second thing architects must do is learn AI by taking an AI course or start working with these new capabilities as soon as possible. This has to be a priority for architects to prepare better for future changes. Finally, architects should begin experimenting with using these capabilities to analyze problems, figure out how to connect these capabilities to datasets and start using them to answer the questions that stakeholders are asking.  The tools are available today to make the shift to the new way of doing architecture analysis.  If you don’t grab your surfboard and start paddling, you will soon be pummeled by the oncoming wave. 

The Future of Enterprise Architecture Analysis – Final Thoughts

Far from threatening the architect’s role, embracing AI can empower and take architects to the next level, providing robust capabilities to help them analyze more efficiently and empowering decision-makers with more insights faster, making organizations more agile in an ever-changing market. If you are a leader of an architecture team and would like to discuss how this trend will affect your organization, Sparx Services North America can help.  Contact us today to get started.

Recent Posts

architecture platform TCO
Architecture Platform TCO
Enterprise Architect Shortage
Application Portfolio Managment
Enterprise Architecture Principles
Business strategy

Learn More

To learn more about the Sparx Architecture Platform and services available from Sparx Services North America…