Architecture is no longer an internal IT concern. In a world where decisions are increasingly automated and scaled through AI, architecture has become a leadership responsibility. This declaration is an attempt to articulate the choices that responsibility entails.
By: Per Imer, CEO, Homerunner
Contains: 1050 words
For many years, architecture was treated as a technical matter. Something that could be optimized, consolidated, and closed off in order to create control, stability, and efficiency. In a pre-AI world, that approach was often sufficient. In a world where decisions are increasingly automated and scaled through artificial intelligence, it no longer is. Today, architecture is leadership, and the choices made around data, openness, and connectedness are, in practice, choices about responsibility.
Closed data is therefore no longer a technical decision. It is a leadership failure. When organizations choose to lock their data away, they are not reducing complexity, but visibility. And when AI is layered on top of a closed architecture, that blindness is amplified. AI is not neutral. It amplifies the architecture it operates within. When the world is shut out, systems become faster at making narrow decisions without understanding the broader consequences.
This challenge becomes especially clear in recursive AI systems, where models learn from their own decisions. Without access to open data and external context, such systems risk collapsing inward, reinforcing their own assumptions and amplifying errors at scale. Recursion without openness is not learning, but self-deception. Technically sophisticated, yet blind to the world beyond its own decision loop. This is an architectural choice, not a distant future risk.
This challenge becomes especially clear in recursive AI systems, where models learn from their own decisions. Without access to open data and external context, such systems risk collapsing inward, reinforcing their own assumptions and amplifying errors at scale. Recursion without openness is not learning, but self-deception. Technically sophisticated, yet blind to the world beyond its own decision loop. This is an architectural choice, not a distant future risk.
At the same time, many organizations have spent disproportionate strategic energy rebuilding systems that have effectively become infrastructure. ERP, WMS, and CRM systems are essential to operating a business, but they are rarely sources of true competitive advantage. Optimizing or rebuilding them delivers limited strategic value if the organization ignores the reality that the majority of value and cost in B2B lies in operations, exceptions, and decisions.
SaaS licenses often represent only a small fraction of the total cost base. The real complexity lives in the physical world, where responsibility, coordination, and consequences are handled every day.
This is where AI should be directed. Not toward marginal license savings, but toward the decision layers where humans currently spend time evaluating, balancing, and taking responsibility. In this context, Logistics as a Service is not about software. It is about responsibility at scale.
It is about orchestrating decisions across systems, actors, and flows, rather than optimizing locally within narrow system boundaries. Orchestration consistently outperforms local optimization, because reality itself is not modular.
Middleware plays a critical role here. Not merely as a technical connector, but as an ethical choice disguised as architecture. Middleware creates the space between action and consequence where context can be assembled, decisions can be evaluated, and judgment can emerge. Without this space, organizations risk building systems that act faster than humans can understand or take responsibility for the outcomes.
This applies across disciplines. B2B and omnichannel operations are fundamentally different in their operational nature, yet they share the same architectural requirements. A single decision backbone that can weigh context differently depending on the situation, but that carries the same responsibility throughout. Fragmented system landscapes produce fragmented decisions. Coherence produces accountability.
The future belongs to organizations that open the world to their systems before those systems begin acting on their behalf. Not because it is easier, but because the alternative is decisions made without visibility and without responsibility.
The question is not whether organizations will eventually lose oversight. The question is when, and whether the right architectural choices have been made before that moment arrives. The decisions made in architecture today are the decisions systems will make tomorrow, on our behalf. And responsibility for those choices does not lie with the technology. It lies with leadership.
1. Closed data is no longer a technical choice — it’s a leadership failure.
2. AI amplifies architecture: close the world off, and you amplify blindness.
3. AI without context is more dangerous than no AI at all.
4. Faster decisions without wholeness only produce faster mistakes.
5. Context is the new intelligence.
6. Large context windows and recursion beat “smarter” models with no visibility.
7. Recursive AI requires open data, otherwise it just accelerates self-deception.
8. ERP, WMS, and CRM are infrastructure, not competitive advantage.
9. SaaS spend is 8–12%; the remaining 90% lives in operations and decisions.
10. The future belongs to those who open the world to their systems, before the systems act on their behalf.
