Eleven services, one runtime.

Inheritance

Hexion's operations team had inherited eleven internal services across a decade of acquisitions. Each had its own scheduler, its own reconciliation logic, and its own on-call rotation. None of them shared a vocabulary.

By the time we were brought in, the team had spent three quarters trying to consolidate them into a single platform. The first attempt had failed because the new system was being built around an idealised view of operations, not the actual one.

What we did

We began with the inventory. Six weeks of structured interviews, system reads, and traffic analysis across every one of the eleven services. The output was a written reading of the environment, including the unofficial workflows nobody had documented.

From there, we deployed Anillion's runtime against the three highest-volume workflows first — port arrival reconciliation, fuel ledger close, and crew rotation. Each ran in shadow mode for thirty days before any cutover.

The remaining eight services were absorbed in a structured cadence over the following eighteen months. No big-bang migration. No production freeze.

What changed

• +Reconciliation moved from a six-week audit cycle to a four-day continuous cycle.
• +Eleven on-call rotations collapsed into two.
• +Reliability against business SLOs reached 99.997%, sustained over eighteen months.
• +The audit packet for the most recent annual review was produced in a single afternoon.

The quiet part

The change Hexion's leadership describes most often is not on the metrics page. It is the absence of weekend pages. The team does not think about reconciliation anymore. The runtime does it.