Operations | Monitoring | ITSM | DevOps | Cloud

Clover’s platform handles more than just payments: inventory, employee management, online sales, and customer loyalty programs are all running on a single monolith called the Clover Operating System (COS). Releasing updates to that platform reliably and without disrupting merchants is one of the hardest operational problems a platform team can face. For a decade, Clover ran HAProxy at the center of its infrastructure.

Legacy infrastructure was not designed for the requirements of the AI era. While large-scale model training remains centralized in data centers, test-time inference is rapidly shifting to the edge to reduce latency and bandwidth consumption. This shift creates a new frontier for enterprise AI, but deploying at the edge introduces significant manual complexity, interoperability issues, and security vulnerabilities.

The telco industry is undergoing a fundamental change. Over the past few years, the increasing maturity of cloud-native infrastructure has accelerated the movement from manually operated and hardware-centric systems to automated, software-defined platforms. Underpinning this change are open source initiatives such as the Sylva project. Sylva is hosted by Linux Foundation Europe and heavily backed by major telecom operators and vendors.

You can already split AI spend by team and by model. But that’s not what your CEO asks in the QBR. The question is what you got for it: what did it cost to ship that feature, to launch that campaign, to serve that customer. And is the AI bet behind it paying off? Now you can allocate AI spend to the outcomes you own: customer, product, feature, the strategic bet on the P&L. Not just the team that spent it.

Every enterprise contract I’ve signed in the last two years has the same clause. “Vendor will not use Customer Data to train machine learning models.” Sometimes it’s a paragraph. Sometimes it’s a whole section. The language varies but the intent is identical: don’t feed our production data into your AI. I get it. I sign the same clause as a vendor. But here’s what’s been bothering me: that clause is a promise, not an architecture.