For a long time, enterprise technology strategy has been built around a simple idea: centralize everything.
One ERP system at the heart of the organization. More functionality pulled into the core. More control through customization.
But as we move toward 2026, that model is quietly — and irreversibly — changing.
What I’m seeing across large enterprises isn’t a rush toward bigger systems or more complex platforms. It’s a shift toward smarter connections, cleaner cores, and architectures designed for constant change rather than long-term stability.
The Move from Monoliths to a Digital Mesh
One of the most significant changes underway is the move away from centralized ERP landscapes toward what I describe as a digital mesh.
In practical terms, this means keeping a clean, stable ERP core — whether ECC or S/4HANA — and connecting other applications around it. Instead of forcing every capability into SAP, organizations are integrating specialist systems through APIs, allowing each component to do what it does best.
This approach aligns closely with SAP’s own direction. The push toward clean core isn’t about limiting innovation — it’s about enabling it without creating technical debt.
This transition didn’t start in 2026. It started years earlier. But what’s different now is momentum.
By 2026, more organizations will feel confident that decentralization doesn’t mean loss of control. In fact, it often delivers the opposite: greater resilience, faster innovation, and systems that can evolve without breaking everything else.
Clean Core as a Foundation, Not a Constraint
In a digital mesh, ERP doesn’t disappear — it becomes more focused.
The core system remains responsible for financial integrity and critical business processes. Innovation happens around it, not inside it.
Applications from partners, suppliers, and internal teams connect at the edges and participate in shared processes without bloating the core. That’s how you avoid the 80-terabyte ERP landscapes of the past — and why modern SAP environments look very different from what we were running even a decade ago.
A key enabler here is the shift toward asynchronous APIs.
Traditional integrations wait for every system to respond before moving forward. Asynchronous models allow processes to continue independently, picking up responses when they’re ready. This becomes essential as landscapes grow more distributed and organizations rely on systems they don’t fully control.
It’s a subtle change, but it makes decentralized architectures viable at enterprise scale.
Where AI Will — and Won’t — Deliver Value in 2026
AI dominates almost every technology conversation, and it will continue to attract investment. But I think expectations — particularly around agentic AI — need to be more grounded.
So far, most AI benefits have been realized inside IT teams. Development cycles are shorter. Testing is faster. Code is written differently.
Useful improvements, but largely invisible to the business.
That changes in 2026.
The next wave of value comes when AI starts to reduce friction for business users, not when it tries to replace core systems or make complex technical decisions.
From System Interfaces to Natural Language
One of the best examples I’ve seen recently was a simple demonstration around purchase orders.
Today, creating a purchase order still involves navigating multiple screens and manually entering data — a process that, in essence, hasn’t changed since the 1970s.
In the near future, that interaction looks very different.
Instead of clicking through screens, a user simply asks:
“Create a purchase order for this vendor and this product.”
The AI handles the structure, validation, and execution behind the scenes.
The underlying system doesn’t change.
The user experience does.
This is where agentic AI makes sense in the short term — acting as an intelligent interface between people and systems. It’s particularly powerful in areas like procurement, supply chain, and operations, where summarization, automation, and natural language interaction remove friction without introducing risk.
Security Is Approaching Its Own Inflection Point
Another area where I see a quiet but important shift coming is security.
Today, most confidentiality still depends on passwords — even when layered with multi-factor authentication. As computing power increases, that model becomes harder to defend.
This is where confidential computing starts to matter.
Rather than relying on credentials, confidential computing creates trusted environments at the hardware level, allowing systems to verify each other without exposing sensitive data. It’s still early, and most ERP environments won’t feel the impact immediately.
But in sectors like healthcare, autonomous systems, and high-tech manufacturing, real-world applications are starting to emerge. By 2026, this moves from research papers into practice.
The Bigger Picture
Across architecture, AI, and security, the same pattern keeps emerging.
We are moving away from tightly controlled, centralized systems toward distributed ecosystems designed to adapt. Not chaotic. Not uncontrolled. But deliberately decentralized.
The organizations that succeed in 2026 won’t be the ones chasing every new technology trend. They’ll be the ones making thoughtful decisions about what belongs in the core, what belongs at the edges, and how systems — and people — interact.
The future of enterprise technology isn’t one big system.
It’s a mesh — and it’s already here.
