One of many network-operations use cases. This example reconciles traffic, topology, and demand data so every capex dollar lands where it earns its return, as one governed, auditable recommendation.
Proves these Lumina features
Network capex has peaked and is now flat to declining. Operators must shift from blanket coverage to capacity where the return justifies it: slicing, fixed wireless access, and private 5G. With multi-billion-dollar build programs, every dollar has to be defensible.
The problem is that the inputs live apart. Traffic and topology sit in the network, ROIC sits in finance, and the enterprise-sales pipeline sits in yet another system. Planning spreadsheets and single-team models cannot reconcile them, so build decisions are made on partial pictures.
Traditional approaches were not built for the scale and complexity of modern operations.
Planning models cannot join network demand, finance ROIC, and the enterprise-sales pipeline, so the build case is always missing a side.
A network-only model optimizes coverage. A finance-only model optimizes cost. Neither produces a decision the other will trust.
When a multi-billion-dollar build rests on a forecast no one can audit, the number is hard to defend to a board.
Which past builds actually paid back, and which did not, is rarely fed forward into the next planning cycle.
Three layers of intelligence working together: reasoning agents, proactive detection, and multi-agent deliberation.
Agents reconcile traffic, topology, and demand across the systems they live in, and produce a forecast where every figure traces to a query against your data. The build case stops being three disconnected models.
Every calculation is SQL you can verify. No black box.
Evaluating capacity build candidates (region: West)... SITE-44 cluster: traffic at 87% PRB utilization, growing 4%/mo Demand: 2 enterprise FWA deals in pipeline ($1.1M ARR) cover this footprint Topology: existing fibre backhaul, no new transport capex needed → Strong build case: high utilization + committed demand + low marginal cost
CONGESTION FORMING: 6 clusters trending past 90% PRB utilization within 2 quarters OVERBUILT: 3 sites below 20% utilization for 12 months, candidates for deferral → Reallocate planned spend from underused sites toward the binding clusters
The Radar scans utilization and demand trends across the footprint to flag where capacity will bind before it degrades service, and where spend is going to underused ground. It points planning at the sites that matter.
The Radar surfaces issues the operator didn't know to look for. Before they become incidents.
Network, finance, and sales agents debate each candidate build: demand pressure, ROIC, and committed pipeline. The output is one governed, auditable investment recommendation with human veto, defensible to a board.
The output is grounded in facts (SQL results), not hallucination. Every recommendation carries a full audit trail.
Network Agent: SITE-44 cluster at 87% utilization, build now Finance Agent: ROIC clears the hurdle only with the enterprise deals attached Sales Agent: both FWA deals are contracted, not speculative Cortex: a similar utilization-plus-committed-demand build last year paid back in 14 months → Recommendation: fund SITE-44 cluster, defer the 3 underused sites. Planning committee approves.
Lumina addresses the four strategic problems that hold operators back.
Evaluate every candidate build against network, finance, and demand data together, instead of one team’s spreadsheet at a time.
Direct capex to where utilization and committed demand justify it, and defer overbuild, in a flat-to-declining capex environment where allocation is everything.
Cortex remembers which past builds actually paid back, so every planning cycle gets smarter instead of starting from a blank model.
Every forecast figure traces to a query, and the investment recommendation carries its full rationale, defensible to a board and a regulator.
Specialized AI agents that power this workflow.
Capacity Planner
Reconciles traffic, topology, and demand into decision-ready build forecasts.
Finance Reasoner
Tests each candidate build against return hurdles and committed pipeline.
Demand & RAN Analyst
Connects enterprise pipeline and FWA demand to the network footprint that would serve it.
This is just one of many use cases
5G capacity & capex planning is one example of how Lumina reasons on operational data. Across Telecom, every domain has use cases where AI agents can add value.
Not ready to commit? Stay up to date as we release new capabilities and industry-specific agents.
View all Telecom use cases