The challenge
A managed service provider is paid to keep other companies' sites up, and judged on one number above all: MTTR. At 16,000 sites, most of that time went before the repair began. The NOC had to read the alarms, work out what had failed, decide who should go, and tell them what to bring.
Dispatch was built by hand, ticket by ticket. Field teams often reached the site not knowing the fault, found they needed a different spare or skill, and had to come back. Every repeat visit counted against the contract.
The solution
Fault management that starts with the cause
Related alarms are correlated into one fault with a probable cause and the alarms behind it. The NOC works a list of faults, not a stream of alarms, and knows what has failed before anyone travels.
AI-enabled dispatch, approved by a person
From each fault the platform proposes the work order, the field team and the priority. A dispatcher approves or changes the proposal and the reason is recorded, so the decision stays with the people accountable for the contract.
Field teams that arrive prepared
The work order carries the fault and the site data. The technician fills in the job form on site and submits photos, and a reviewer confirms the repair before the job is closed.
One way of working across every site
The stages, approvals and dispatch rules of each type of job are set once and applied everywhere, so a repair runs the same way whichever team or region picks it up.
The results
MTTR came down by 35% across the 16,000 sites, because the hours once spent reading alarms and deciding who should go now go on the repair.
Seven in ten jobs are now dispatched with AI assistance: the work order, the field team and the priority are proposed, and a dispatcher approves.
Field teams arrive knowing what failed, and each job is closed on a reviewed job form and photos the provider can show its customers.
The takeaway
Most of the repair time is spent before the repair starts. Preparing the diagnosis and the dispatch, and leaving the decision with a person, is where a service provider wins that time back.