Business Architecture Is Getting a Second Wind, and AI Is the Reason Why.

Business architecture has been around for a while. It emerged in the late 1980s largely as a way of defining business structure in Information Systems Management, with Edwin Tozer espousing it and information management in 1986, and then William R. Synnott provided a framework in 1987. It has percolated along, with OMG touting it in 2010, The Open Group Architecture Framework (TOGAF) promoting it, ArchiMate providing a metamodel for it, and others promoting it as well. In the late oughts Bodine and Hilty declared it an “an important new corporate activity” (Business Architecture: An Emerging Profession.

Created by ComfyUI, commercial trade in the style of Ukiyo-e.

In most places that I have worked, I have not seen strong business architecture practices. A 2024 study has shown the 84.1% of organizations do not have a mature business architecture practice — despite being around as a concept for almost 40 years, and being promoted as a major concept of enterprise architecture for more than 20 years.

The reasons for this are diverse, but include:

• Where does it belong — Architecture is a historically a technical endeavor, coming from networking, application and data processing professionals. Business architecture crosses over into a world that traditionally does not understand and sometines does not value technology. When tech folks start talking about business architecture, the busibess guys may see it as the barbarians storming the gates. They may think the structured approach is too heavy and does not provide value. Business people, on the other hand, rarely initiate business architecture initiatives. They just are not exposed to it and do not come across it in their day-to-day business management processes. All this leaves business architecture out in the cold.
• It’s too much effort — Immediate concerns are taking all the air in the room. I have worked many places where executives are so busy trying to fix problems, that they are not getting ahead of the problems. Delegate people! An executive should not be in the weeds. They should be focused on defining the needs, finding great talent, providing strategic direction (i.e., making sure strategy is being executed) and, perhaps most importantly, shaping the culture of the organization. You may have been good as a manager, but a leader has other responsibilities. Some who get promoted, never understand that.
• It requires discipline — Many organizations just have not developed the organizational discipline to develop business architecture. This is not helped by the issue that there are divergent practices, a lack of business architecture training, and poor tooling that make it harder to implement. With the pace of business change, this only compounds the issue. Disorganization, lack of tooling and poorly defined practices result in even more work to keep up with the changes. Without the tools in place, for many executives it is easier and more expedient to rely on heuristics rather than use data-driven decision-making.
• Executive buy-in is hard to get — It can be difficult to explain the value proposition of business architecture. There may be slow time-to-value because the base processes have not been put into place, and executives are prone to over-estimating their capabilities. They rely on heuristics, their experience, intuition. This kind of hubris can lead to subpar performance.

I believe the business architecture does have value and is important for organizations to include in their enterprise architecture practice. The top goals for business architecture are:

1. Align strategy with execution — IT is an expensive endeavor. With putting so much resources into it, it should be moving the ball forward. If it is not executing on strategy, that is a huge problem.
2. Provide a path to the future— Defining this path is the first step towards strategy execution. It is defining what the business needs to do in order to execute the strategies. This can be divided into two steps: Identify the current state of the business capabilities and then map the future capabilities.
3. Enable Better Decision Making — Business architecture creates catalogs of information that can be used to gain a holistic view of processes, define relationships between applications and business processes and roles, and identify value streams. With data available, decisions can be made with greater confidence, and more quickly.
4. Improve strategic planning — Business architecture can show dependencies, gaps and overlaps, helping leaders understand where savings can be derived and where to invest.
5. Consolidate and structure the organization — With the data at hand, a unified view of the organization is available. Silos can be broken down, execution aligned, and downstream decision-making is based on true knowledge of the organization and where it is headed.
6. Provide a path forward — Mapping the current and future states, and sharing those maps with team members, you provide an idea of what it is the organization is trying to achieve and how it is trying to achieve it.

While I am a believer in business architecture, I am not the only one. Gartner’s Hype Cycle for Enterprise Architecture, 2025 has Business Architecture as an important aspect of Enterprise Architecture, with high value to the organization and becoming mainstream within the next 5 years.

Get Jake’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

Time will tell if it does become mainstream, but I think there is one important innovation that may help it along: Artificial Intelligence.

Artificial intelligence won’t directly overcome the issues with where business architecture belongs, or executive buy-in, but it will help with the discipline and effort issues.

As AI builds steam, it will become clear that disorder is a hindrance. Agents will be just as confused as humans when faced with disorder. In order to make use of agents and reduce the risk associated with their autonomous actions, the enterprise will need to get its house in order. Technologically, this means understanding the current state across the architectural domains, including the business domain.

Technical architectures will also change. Monolithic applications, which lock up functionality internally, will be replaced with composable architectures. This means event-driven architectures, microservices and orchestrations. These are the ways that AI will work in the future. Agents will both raise and respond to events as well as microservices. Orchestrations will enable more complex, long running, multi-agentic workflows. All of this needs to be aligned with the business. This means business processes, capabilities and strategy — Business Architecture elements.

The business world increasingly requires more agility, more coordination and more strategic execution. AI agents will power this, but it will require information. This is especially important when it comes to agentic software development. As the human steps away from the software development, all of the intuitive knowledge and all of the historical context steps away as well. Humans retain knowledge, however imperfectly, over a long period of time. Agents are created, executed and are destroyed. Their context timeline is very short. The human may know that the application is an accounts payable system, the agent must be told this. The human will know at least some of the history of the accounting systems internally and what changes occurred over time, the agent may not. The human will know how it will integrate with an ERP system, or third party accounting software or how invoices will be coming in, the agent will not. This is especially important as the requirements and specifications begin to be developed by agents.

This is where business architecture comes in. The first important thing is to catalog the necessary information. Business Processes, organizational structure, and roles are the initial priority — these can be used by agents who are food sped requirements from humans. Later, business capabilities, value streams, mission and objectives will be important. These all provide critical information that allows the agents to live up to their autonomous nature when helping define the formal requirements and specifications that will be used by the programming agents.

Business architecture provides the following benefits to AI Agents:

• It gives agents a shared understanding of why the system exists. Like with humans, this provides context for downstream decision making.
• It provides a blueprint for what the current business environment looks like — what are the capabilities and what maturity level they are being set to, where do handoffs occur, what processes can be automated and to what extent, and what constraints are put on the agent’s decision-making.
• It provides the interfaces between humans, systems and agents — allowing agents to better understand what data they can access, and where humans fall in the process.
• It provides structure data, including data models, that agents will use, including in support for reasoning frameworks and Retrieval-Augmented Generation (RAG).
• It accelerates development by aligning the agent with business priorities — answering the question of where does the agent need to put more effort based on the capability roadmap, or defined high-value capabilities, for example.
• It sets boundaries and better allows agents to identify and mitigate risks by understanding the regulatory and compliance environment, what internal governing processes there are, and what limits have been set on their abilities.

The challenges moving forward, along with the opportunities, are immense, for sure. But it won’t all be at once. I personally believe the real value of business architecture will be realized through agents. It gives the purpose, structure, rules and context they need to be effective digital workers. I think there was value before, but with AI, that value will be even greater.

Much of this knowledge should be stored in a Enterprise Knowledge Graph (EKG). This is a specialized database that includes important architectural information from across the architectural domains, including business. As a graph, it contains both nodes (a point of information) and relationships (how two nodes are associated). For example, the EKG could include applications, business processes, roles. A business process may be executed by one or more roles For the AP example above, it could be an AP Specialist for approving invoices under $1000, or a CFO for invoices over $10,000. A business process may also use applications, such as a third-party purchasing system, or an ERP for business partner lookup.

While AI will expose the value of business architecture, business architecture instantiation will be enabled by both automations and AI. Putting into place crawlers, ai agents and a structured eventing system will help populate the EKG with business architecture knowledge. Scanning documents for business processes is a good first step to populating the business architecture catalog. Early on, it will still need a human eye to validate the catalog, but it will help speed the process along. Later, as more structure business processes are developed, it will keep that information updated in a timely manner. Agents will both make use of, and keep order in the EKG, including business architecture data.

As organizations step into this new era, the convergence of business architecture and AI isn’t just a technical evolution — it’s a strategic imperative. The companies that thrive will be those that recognize that intelligent agents are only as effective as the structure, clarity and intent we give them. By investing in business architecture now, enterprises lay the groundwork for AI systems that are not only powerful, but purposeful. The second wind for business architecture is already here, carried by the accelerating force of AI, and the organizations that embrace this pairing will be the ones best positioned to navigate — and shape — the future.