In the new book Ecosystems Architecture: New Thinking for Practitioners in the Age of AI, authors Philip Tetlow, Neal Fishman, Paul Homan, and Rahul share that business complexity has increased to the point where traditional systems approaches no longer work. This clearly was foreseen by futurist Alvin Toffler, author of Future Shock and The Adaptive Corporation. In 1985, Toffler noted that the corporate environment is changing so swiftly and fundamentally “that the structures designed for success in any industrial environment are almost inappropriate today.” Proving this point is public company longevity, which has decreased from an average of 33 years in 1960 to 10 years during the COVID-19 crisis.

Foundations

According to Tetlow et al., the Industrial Revolution started with the movement from small-scale artisan endeavor to widespread industrialization. This started an upward spiral into globalization and eventually an ever-increasing network of electronic communications. Over the last 50 years, we have gone from mechanization to electronic circuits to an information economy.

Klaus Schwab of the World Economic Forum suggests that “the fusion of digital technologies … and their interaction across the physical, digital, and biological domains … make the Fourth Industrial Revolution fundamentally different than previous revolutions.” Tetlow et al. add that complexity has increased as information has become networked and the heart of businesses.

In the 1980s, systems thinking came to the forefront. A progenitor of this approach was organizational theorist Russell Ackoff. In Creating the Corporate Future, Ackoff claimed that as the rate of change increases, the complexity of problems that face us also increase. “The more complex these problems are, the more time it takes to solve them. The more the rate of change increases, the more the problems that face us change and the shorter the life of the solutions we find for them,” he said. Ackoff then proclaimed a post-industrial age and stressed the importance of systems thinking and synthesis versus analysis.

Without question, Tetlow et al. are correct in suggesting that the increasing scale and complexity of systems brought about enterprise architecture as a function a decade or so later. However, today, they say that complexity has increased to a level where it is increasingly difficult to intuit or model how a system works.

The problem is that the architects of change — enterprise architects — need to contain the problems to solve them. This no longer works as systems complexity increases. However, as the authors share, software engineer Grady Booch foresees a third golden age of architectural thinking, and this will power us forward to support more complex organizations and ecosystems.

In this final era is architecture driven by generative artificial intelligence (GenAI), large language models, and neural nets. Here, the authors see human experts partially being replaced by machines; you can consider it augmentation by synthetic assistants.

The Advent of Ecosystems Architecture

In the 1980s, many business leaders began to breach the boundaries of traditional businesses and started to wrestle with the complexity of connected communications between systems of systems. This, accordingly, prompted talk of “systems of systems.” Ackoff called these problems “a mess,” stating that a mess, like any system, has properties that none of the parts have. These properties are lost when the system is taken apart. According to Ackoff, “The solution to a mess depends on how the solutions to its parts interact.”

To formally address the systems of systems mess, ecosystems architecture emerged between 2014–2019. According to Tetlow et al., this led to the concept of “hyper-enterprise connectivity.” This concept recognizes that most enterprises need to manage an ecosystem — a complex web of interdependent relationships aimed at creating and allocating business value. For modern organizations, this represents an extended network of enterprises that exchange products or services within an environment.

An important concept introduced by the authors at this point is that business organizations should not restrict how, where, or what commerce takes place. An ecosystem can be broad by nature, potentially spanning multiple geographies and industries. This leads to systems of cooperation and orchestrations. Here, value is generated through external coordination and collaboration.

Peter Weill and Stephanie Woerner from the MIT Center for Information Systems Research elaborate on this concept in their book, What’s Your Digital Business Model. They share how organizations with great customer experience will be naturally attracted to being “ecosystems drivers.” These firms are organizers of an ecosystem, a coordinated network of enterprises, devices, and customers. The ecosystems create value for all participants. To make this all work, ecosystem drivers need to establish digital ecosystems. To achieve this, Tetlow et al. claim that enterprises need to work to remove organizational boundaries between digital and physical channels. This means technologies must grow to be more distributed, powerful, and intelligent. Without question, AI can help mask the growing complexity.

Pulling Apart Architecture

This, to a certain extent, pulls apart enterprise architecture from ecosystem architecture. Instead of functions, attributes, and contained-within components, ecosystem architects abstract nodes and correlate them with clusters, domains, regions, communities, colonies, enterprises, systems, components, aspects, attributes, or whatever levels of abstraction are needed. To be clear, Tetlow et al. see ecosystems architecture as an additive discipline. The problem, however, is with ecosystems complexities and the scale of the problems; you might as well mask any systems’ structure up front. Here, ecosystems architects look at cumulative contributions of business functions.

To squeeze out structure, ecosystem architects need to understand that integration and interoperability must rise in prominence. This is especially the case during the early stages of architectural thinking. Ecosystems architects must also understand how an enterprise is constructed. Here, each silo becomes a node, and nodes become a container for other nodes.

Without question, a modern enterprise blends dependent and independent business units that coexist within complex ecosystems. Invariably, ecosystems can transcend corporate boundaries: where does an enterprise end and open-ended ecosystem thinking begin?

What Has Changed

Historically, leaders only worried about getting the best from their part of the value chain. Today, change has become the only real constant in business. This means that more comprehensive approaches are needed.

Complexity can no longer be “boiled out.” For this reason, brute force or simple models no longer work. Architecture needs to deliver systems that can tolerate uncertainty and rapid change. According to Jeanne W. Ross, Cynthia M. Beath, and Martin Mocker (authors of Designed for Digital), to succeed digitally, “companies must define data, business, and infrastructure components and design them for reuse.” To do this well, Tetlow et al. suggest that architects need to deeply understand the mechanics at play in the world around them. Doing so involves broadening skills, better understanding complexity, and ensuring that IT architecture does not become a blur.

Tooling for Ecosystems Architecture

Architecture today often consists of boxes and lines. In its simplest form, these represent inputs and outputs. Tetlow et al. ask whether simple abstraction and modeling still work as complexity increases. Model-driven architecture worked because it was graphical and intentionally reductionist in nature.

The problem is the world is not that simple, and simple models, while elegant, may not work. The authors claim that, fortunately, today’s AI does not need to reduce or separate the complexity. AI can help architects with the complex mix of detail and enable organizations to achieve hyper-enterprise scale.

To educate architects and lay people, Tetlow et al. describe, in a nontechnical way, how new technologies transform architecture. They walk readers through the details and applications of graph theory, vectors, cosine similarity, and GenAI. As the authors explain, AI enables models of ecosystems architecture that do not abstract away the complexity. Because of these new technologies, the authors contend that tools and techniques are within our reach to perform ecosystems architecture and deal with complexity. AI-infused architecture eliminates the need of architects to synthesize the scale and complexity playing out behind the scenes. We clearly are at an interesting tipping point. As the pace of change accelerates in areas like GenAI, it is becoming increasingly plausible to use synthetic help to augment and automate architectural practice.

Parting Thoughts

Change clearly is not new. It is part of the human condition. In Tales of a Traveller, Washington Irving wrote, “There is a certain relief in change, even though it be from bad to worse! As I have found in traveling in a stagecoach, that it is often a comfort to shift one’s position and be bruised in a new place.” Our “new place” is the pace of change and the complexity that it creates. As we move into a hyperconnected economy, the complexity will clearly outstrip our ability to model and connect effectively. The opportunity today is to use AI to manage that complexity to build the enterprises of the future. This makes Ecosystems Architecture a worthy read for enterprise architects and digital business strategists.