Two teams spent months building the same integration because two project managers each carried the same request to a different team.
The demo that stopped the room
At a large medical technology company, a development team opened its sprint review with a demo. One of the developers pulled up a connector his team had built to move data between two of the company's business systems. The business had been asking for this integration for most of a year, and now it was working on screen.
Across the table, a developer from another team stopped typing. She asked him to explain again exactly what it did. He did, more slowly. Then she turned her laptop around. Her screen showed the same connection, unfinished. Her team had been working on it for months.
He looked from his screen to hers. His team had started later, he said, and had finished in about four weeks. Her team lead leaned in and asked the question everyone was now thinking: who told you to build this? He named a project manager. The team lead turned to her own developer, who named a different one. One of the delivery leaders in the room said it out loud: they had wasted development time they could not afford to lose.
Two requests, two doors, no shared view
Nobody in that room had done anything careless. Two project managers had each taken a request from the business straight to a development team. The requests were almost identical. The project managers hadn't coordinated, and once the work was with the teams, neither team had any way to see what the other was doing.
It wasn't a one-time slip. In this organization, a project office had long stood between the business and the development teams, and its project managers routinely handed work directly to developers. Business partners were committing to delivery dates without knowing how much capacity the teams had. Work was scattered across the company's work management tool, so there was no single place to see what was already underway before something new started.
Given that structure, the duplicate was predictable. Each handoff made sense locally. Nothing on the path forced the two requests to meet.
-
Business needs integration Asked for over most of a year
-
Where it broke PM one briefs team one Straight to the developers
-
Where it broke PM two briefs team two Near-identical request
-
Where it broke Both teams build unseen No shared view of work
-
The reveal Overlap found at review After the build was done
The problem was not downstream
An earlier effort had tried to fix the development teams and failed. The teams were doing exactly what they were told. The problem sat upstream, in who controlled the space between the business and the developers.
The duplicate connector showed that conclusion clearly. Telling project managers to communicate better, or asking developers to check with other teams before they started, would just have added another informal habit to a path that already depended on them. The path itself had to change.
That meant a decision about authority, not process. Senior leadership made the change explicit and visible. The project office kept a real voice in deciding what got built and in what order. What it lost was the ability to give work to the developers, or reorder their priorities, outside the agreed path.

One way in, one place to look
Three changes in how work reached the teams addressed this kind of duplication directly. First, a single weekly forum brought business partners, the project office, and development leads together to review incoming demand, agree on priorities, and move work into development through a defined intake, with explicit definitions of ready and done. Demand from different parts of the company fed into that forum through one workflow, and no group could bypass it on its own.
Second, the scattered work was consolidated. Delivery work that had been spread across the work management tool was moved into a single project space, and each team got a board with the same structure as the others.
Third, large business outcomes were broken down by a standing refinement group into pieces of work that took one to three months, before anything was handed to a team. A business leader matching work to teams could now see which pieces were already in progress before adding new ones. That is the check that catches a second connector before anyone starts building it.
- PMs hand work straight to teams
- Business commits without capacity
- Work scattered across the tool
- No view of what is in flight
- One weekly forum reviews demand
- Work enters through one intake
- One project space, common boards
- In-flight work visible first
What a single view made possible
The relationships didn't transform overnight, and how the project office came to accept its new seat is a story of its own. What changed was the system around them: once work couldn't be reordered without the change being visible, the quiet reordering stopped.
For the business leaders, the change showed up in what they could see. A leader could open an outcome, see the pieces of work beneath it, and see which teams were working on them. Before committing new work, they could see what was already in progress.
This one doesn't end with a tidy number for the rework the discovery saved. What it does show is what produced the duplicate in the first place. Two capable teams built the same thing because the organization let requests reach them without passing through anything shared. Closing that gap was a structural decision, and the structure, not anyone's goodwill, is what keeps the next duplicate from starting.
Duplicate work is rarely a team failure. It is what happens when a request can reach a team without passing through anything shared.
Share With Your Team
Choose who it is for. You send it yourself from your own email; Agileon never writes to your colleagues.


