Synopsis
- Thousands of stalled jobs surfaced and unblocked in the initial production sweep
- Training time for new coordinators reduced from weeks to days
- Sharp drop in "what's next?" and "why is this stuck?" support tickets
- Coordinator capacity reclaimed — less time chasing status, more time managing exceptions
- Foundation laid for ongoing process refinement as user feedback feeds the model
Article
When the data moves but the people don't
The carrier's construction workflow was, on paper, automated. Data flowed in from multiple systems. Milestones triggered the next phase — business approval, funding approval, permitting, build, closeout. The system knew when a job had advanced.
The problem was that the system wasn't telling anyone.
Construction coordinators — each responsible for dozens, sometimes hundreds of active jobs — had no notification when a job moved into their court. They tracked work in spreadsheets and home-grown Access databases, and pieced together status from email threads and free-text comments. The systems didn't talk to each other, so neither did the people using them.
The cost wasn't just slowness. It was invisible work-in-progress. Jobs that should have moved forward sat untouched, sometimes for months, because nobody knew it was their turn.
What we built
We mapped the carrier's existing data-driven milestones onto a standard process model in Camunda. Through interviews with coordinators, we learned not just what triggered a phase change but why — and who needed to know, and when, and what they needed to see to act on it.
The process model became the brain. It decides:
- which user owns the next step
- when they should be notified
- which screen, system, or form they should land on to complete it
That logic is delivered to the coordinator through BP3's Brazos Task Manager — a single inbox-style view of every job that needs them, in priority order, with the right context already loaded.
For the first time, the work came to the coordinators. They stopped reverse-engineering status from spreadsheets and started clearing tasks.
What changed after go-live
The first surprise was the backlog. As soon as the process model went live, it identified thousands of jobs that had been silently stalled — held up by issues nobody had flagged, waiting on people who'd never been told. Each one came forward with a clear next step and an owner.
From there, the changes compounded. New jobs were assigned the moment they were ready, not when a coordinator happened to notice. Onboarding new coordinators got dramatically faster, because the system was the training: the right task, on the right screen, with the right context. Support tickets querying "what do I do next?" and "why is this job stuck?" — once a steady drumbeat — dropped to a fraction of their previous volume.
Adoption grew on its own. Once coordinators saw that the inbox was accurate and the work was real, they stopped going back to their spreadsheets.
Why this pattern matters for telecom buildouts
Most carriers running fiber, tower, or small-cell programs have the same shape of problem: the data is fine, the systems are fine, but the handoff between systems and humans is held together with tribal knowledge and email. A process engine plus a task manager turns that handoff from a coordination tax into a routine event.
If your construction or buildout team is running on spreadsheets, status meetings, and Slack pings to figure out what's stuck — that's the workflow we replace.