01 · Legacy OSS modernization
Façade architectures around the system that actually runs the network.
Modern API in front of the legacy core. Reads first, writes when justified. The OSS keeps running; new capability ships from the façade.
Sector · Telecom operators
Delta Telematics builds OSS/BSS modernization, agentic AI on legacy customer systems, and enterprise platform extensions for field operations. We work with Tier-1, Tier-2, and regional operators across Canada. The pattern that brings most telecom engagements to us: a legacy system that still runs the network, a vendor module whose licensing has outpaced its value, or an AI use case that generic chat tools cannot reach. We modernize without rip-and-replace; we extend the platforms you already own without paying for what we can build; we wrap legacy OSS in agentic AI without putting customer data through a foreign cloud.
The single most expensive feature of telecom IT is the OSS that has been running for fifteen years and still does. Replacing it is a multi-year, multi-million-dollar program. Leaving it alone forfeits everything modern AI and orchestration could give the operations team. Most operators are stuck between those positions.
Our position is that the legacy OSS is an asset, not a liability — once you put a façade on it. A custom MCP connector + a Claude Agent SDK runtime turns three decades of accumulated business logic into a queryable, governed AI workspace. Operations teams get natural-language access to the system of record, with every query audited. The legacy system runs unchanged.
The other recurring telecom problem is per-resource vendor licensing on field-operations modules. At a hundred technicians the licensing is annoying. At a thousand it dominates the spend. We have replaced the leading vendor field-service module with a native build on the operator's existing enterprise platform that mirrors its capabilities, costs a fraction of the recurring fee, and lets the operations team configure dispatch logic the vendor does not support.
Where we typically engage
01 · Legacy OSS modernization
Modern API in front of the legacy core. Reads first, writes when justified. The OSS keeps running; new capability ships from the façade.
02 · Agentic AI on the OSS
Claude Agent SDK + custom MCP connector + skill files that capture institutional knowledge. Every query logged and replayable.
03 · Legacy field-service replacement
60–75% reduction in annual field service licensing cost in past engagements. Custom dispatch rules the vendor module did not support.
04 · NOC AI assistance
Reduce mean time to action by giving the on-call engineer a sovereign AI agent that already knows the system layout.
05 · Cross-system event pipelines
Event-driven integration that's small enough to understand, fast enough to scale, and observable enough to debug at 3am.
Patterns
Telecom · Pattern
Native enterprise-platform builds — 60–75% reduction in annual licensing cost, full data residency, workflow logic owned by the operations team.
Telecom · Pattern
Claude Agent SDK + custom SQL connector + reusable skill files turned a legacy customer system into a queryable, auditable AI workspace.
For NOC and OSS teams
Tell us what's running, what's expensive, and what you wish you could ask it in plain English. We'll come back with a fit assessment and a no-rip-and-replace path.
Talk to an engineer