An authority's problem is not running the buses. It is knowing whether the operator did.
A Unified Metropolitan Transport Authority plans, contracts and pays for services somebody else operates. That changes what the software is for: less dispatching, more evidence — what was promised, what ran, what it cost, and what the contract says happens about the difference.
Four things an operator's system will not give you.
One network across operators
Several operators, one network model. A passenger's journey crosses contracts; the plan and the performance figures have to as well.
Planned versus actual, independently
The service specification the authority issued, compared with what the vehicles actually did, computed from data the authority holds rather than from a report the operator writes.
Contract logic that runs
SLA thresholds, penalties and incentives applied to the measured figures, with the calculation shown. A deduction nobody can reproduce is a deduction that gets argued away.
Evidence that survives audit
Management boards and reports built from the same records as the payments, retained so a question six months later has an answer.
Whose data is it.
Most disputes between an authority and an operator are not about performance. They are about whose figure is the real one.
Ingest directly
Position and telemetry taken from the vehicles rather than from a monthly summary, so the authority's figure does not depend on the operator's arithmetic.
Role-based access
Operators see their own operation; the authority sees all of it. One platform, one data model, and no shared login pretending to be governance.
The same definition
On-time performance means one thing across every contract, because it is computed once from configured limits that both sides can read.
What this does not settle.
Start with the contract, not the fleet.
Show us what the operator is measured on and how it is calculated today. That is where an authority deployment either works or does not.