Resources
White papers, demos, and the case study — everything we’ve published on running a carrier network with a smaller team.
Most of what follows is written for people evaluating a change to how their operations run, not for people who have already decided. If you want the product itself, start with the platform or the six agents.
White papers
Killing the Alarms
The architectural argument for why the human-in-the-loop NOC stopped scaling. The data fabric behind sub-second correlation, the specialised-agent model, deployment and resilience — plus an honest section on where the model is hard and where the labour argument gets oversold.
Read it →Tired of Tickets? Ask Frank.
How an AI Tier 1 agent handles the support volume that doesn’t need human judgment — knowledge base questions, incomplete reports, duplicates, and false alarms. Includes the five decisions Frank makes on every ticket and the confidence thresholds behind each one.
Read it →The Last Engineer Who Knew
Network firefighting is no longer a strategy. Three forces — retirement, downsizing, and expert flight — are draining institutional knowledge from telecom faster than it can be replaced, while networks span four to six technology generations at once. Why previous knowledge management attempts failed, what retrieval-augmented systems change, and a phased roadmap that starts delivering in weeks rather than years.
Read it →The Autonomous Carrier
Powering the lights-out NOC. Why the economics of manual operations have stopped working — maintenance costs compounding while service revenue flattens — and what an agentic operating model changes for the customer, the operations executive, and the CFO. Written for leadership rather than for the NOC floor.
Read it →Ending the OSS/BSS Integration Tax
A typical Tier 2 or 3 carrier runs eight to fifteen OSS/BSS platforms, and every vendor change triggers another integration project. Why SI firms, iPaaS middleware, internal teams, and generic AI coding tools all leave the same cost structure intact — and how conversational integration, version control, and outcome verification break the dependency on a specialist at every integration event.
Read it →5G Slicing Demands New OSS Tricks
Service catalogues built for permanent services have nowhere to put a mission that exists for thirty minutes. What slice-aware operations require from topology, correlation, and the order-to-revenue path when the network reshapes itself continuously — the technical depth behind the DTW Ignite Catalyst.
Read it →The Rationalization Window
Turning OSS/BSS tool sprawl into AI-native savings. The case for consolidating the assurance layer rather than adding another platform on top of the six you already run, and how to time that decision against your existing renewal cycles.
Read it →See it running
Recorded walkthroughs against a live FTTH environment — correlation collapsing an alarm storm, Frank working a support queue, the supervisor briefing, and the topology map updating as the network changes. All of them are on the Rapax channel.
If you would rather see it against your own network than a demo one, that is what the fifteen minutes is for.
Case study
A mid-market fibre ISP running GPON and XGS-PON across a multi-vendor footprint had solid subscriber telemetry and no way to answer the only question that matters during an outage: how many customers are down right now.
Rapax consumed their existing Kafka streams, built the topology graph from backbone to subscriber, and published correlated service-impact events back into their stack. No device agents, no schema changes, no downtime, and the initial deployment took under two hours. Reference calls available on request.
Writing
Shorter pieces on network operations, AI in the NOC, and what the shift to agentic operations actually asks of an operator — including the series introducing Frank and the slice-aware operations work shown at TM Forum DTW Ignite.
Run it against your own network
Rapax deploys in five days. You evaluate for thirty against success criteria you define in writing before anything is signed, for $25,000 — credited against the licence if you move forward. If it does not meet them, the evaluation ends and you owe nothing further.
The ask is fifteen minutes, not a demo. Book 15 minutes · sales@rapax.app