The challenge
Every mobile operator has more places to invest than budget. Deciding where to build a new site or switch on 5G means weighing demand, the customers behind it, the capacity already in place and what each site can physically take. Each sat with a different team and in a different system.
The plan was assembled by hand from separate exports. A ranking took weeks, was hard to defend when someone asked why one area came before another, and had to be rebuilt whenever an assumption changed. Network figures alone could not show where customers would feel an upgrade.
The solution
Customer data beside network data
Customer data and data from across the network domains were put on the same page for each site and area, so demand, experience and capacity are read together instead of in three reports.
A predictive model for where demand is going
An AI model learns from usage and capacity to show where the network will come under pressure. Planners see the areas that will need investment, not only the ones that already do.
Options compared on the same data
For each area the options sit side by side, each with its driver: optimise what exists, upgrade capacity, or build a new site.
A ranked plan the planners approve
New sites and 5G rollout are ranked, with the assumptions behind the ranking listed in the plan. Planners change an assumption and see the ranking move. What to build stays their decision.
The results
The AI model behind the ranking reaches 90% accuracy, so the planning team can rely on it when choosing where to build.
Some 1,200 sites are now ranked for new sites and 5G from one plan, built on customer data and network data together.
Producing the plan takes 60% less time than assembling it by hand from separate exports, and it no longer has to be rebuilt when an assumption changes.
The takeaway
A capex plan is only as good as the data under it. With demand, customers and site constraints in one decision, the ranking can be explained, defended and changed in an afternoon.