Aziz Benmalek is CEO of Aeris
Three times in my career, I’ve watched the center of gravity in enterprise technology shift, from on-premise software, to the cloud, to SaaS. I built businesses inside each of those shifts, at Microsoft, Splunk and Sage. Now I run a company at the center of what I believe is the next one, so I have a stake in getting this right: AI is not just eating software. It exposes which layer of your business it actually has to run through.
That layer is your strategic control point. Right now, identifying it is one of the most important exercises a leadership team can undertake.
For some companies, that control point is still software; for others, it’s logistics, payments or hardware. In my industry, it’s the cellular core network we operate and the control plane running on top of it. As AI agents are evolving from answering questions to taking action, the companies closest to these control points will hold a critical advantage.
Intelligence At The Edge
Every month, connected devices carry more computing power, more capability, and increasingly, their own AI. This is physical AI: intelligence moving into the hardware itself. Connected cars are becoming software-defined vehicles, almost mini data centers on wheels. The same shift is underway across utility grids, manufacturing equipment, medical devices, fleet trackers and soon, robots. Small models running on these devices will increasingly collaborate with larger models in the cloud, making decisions together rather than relying on one centralized brain.
At the same time, networks are becoming programmable. Technologies like standalone 5G let enterprises manage and optimize connectivity in ways that weren’t possible before. The network is starting to look like an operating platform, and that changes the role of AI. Instead of simply analyzing information after it’s collected, AI can observe what’s happening, decide, and in the form of agentic AI, act immediately.
Shift From Watching To Acting
Most organizations have spent years investing in observability, monitoring systems, collecting telemetry and building dashboards. But a dashboard is a confession: it says the labor of noticing is yours. In the agentic era, the shift that matters is from observability to decision-making to enforcement. The closer you sit to the layer where decisions are enforced, the more valuable your position becomes.
In connected-device industries, that layer is the network. My company operates the cellular control plane for more than 115 million devices, across 7,000-plus enterprises and over 30 global mobile network partners. That gives us visibility into billions of network events every day. More importantly, it gives us the ability to act.
Imagine a cybersecurity attack aimed at a fleet of connected vehicles. A platform operating at the network layer can redirect communications or trigger a security kill switch to contain the threat, instead of alerting the manufacturer after the damage is done. That’s the difference between documenting an incident and preventing one.
But enforcement power raises a fair question: who is accountable when a system acts on its own? Enterprises are racing to secure “nonhuman identities,” the AI agents and APIs now acting inside networks without clear ownership. I believe the winners of this era will be the platforms that are the most governable, not the most autonomous. Every automated action a network takes should be logged and tracked for threats, changes and incidents, explainable and traceable to a policy that a human established. Autonomy without accountability is a liability waiting to surface.
Find The Layer You Can’t Afford To Rent
Enforcement is only worth having if you’re standing on the right layer to begin with. History shows that success often hinges on controlling one strategic layer of a market. Amazon built a fulfillment network that became a platform. Apple integrated hardware and software so tightly that iOS became inseparable from the iPhone. Nvidia turned a chip business into a platform through CUDA. Each company found a layer everyone else depended on and built value on top of it.
In our case, that means the network, or the control plane: 20 years of infrastructure and carrier integrations that can’t be assembled quickly or rented at any price.
The pattern is the same everywhere: own data no one else has and sit as close as possible to the point where decisions are made. That’s where AI creates the most value, and the companies in the strongest position will be the ones that own the infrastructure, data and control points those models depend on. Every business has a layer it can’t afford to rent. The task now is finding yours, before someone else builds a platform on top of it.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?