The challenge
Energy is the largest running cost of a tower portfolio, and the hardest to track. Grid, DG, battery and fuel were read in different places. A site could waste energy for weeks before anyone noticed, and nobody could say which sites to fix first.
The NOC had the opposite problem: too many alarms. Every site raised its own, and the same fault raised them again and again, so the shift went on clearing duplicates. Field teams were dispatched on thin tickets, and colocation requests waited while people pulled lease, tenant and capacity records from separate systems.
The solution
Energy optimised site by site
Grid, DG, battery and fuel data were put on one page per site and compared with how the site was expected to run. The sites losing the most energy came first, each with the reason and the action to take.
Predictive models for anomalies and fault management
Models learn how each site normally behaves and flag what does not fit before it becomes an outage. Related alarms are correlated into one fault with a probable cause, so the NOC works a list of faults, not a stream of alarms.
AI-assisted workflows for the field
An approved action becomes a work order with the finding attached. The field engineer arrives knowing what failed, fills in the job form on site, and the job is closed only after the form and photos are reviewed.
Faster answers on colocation
Site, lease and tenant records sit beside the capacity left on each tower. A request is checked against what the site already carries, and the assessments still required are listed up front.
The results
Energy use came down by 8% across the portfolio, by finding the sites losing energy and fixing those first.
Productivity rose by 28% across operations and field teams: less time clearing duplicate alarms, fewer visits made blind, and jobs closed first time.
The NOC now handles about five times fewer alarms, because repeats and related alarms arrive as one fault with its probable cause.
Colocation requests are answered from one record instead of three systems, so the commercial team can reply to a tenant while the request is still live.
The takeaway
A tower company earns on two things: what each site costs to run and how many tenants it carries. Energy, faults, field work and capacity in one place move both.