Project Rescue & Stabilization
We adopt half-finished software someone else walked away from, work out what was actually promised, and deliver it properly.
Rescue starts where another engagement ended badly: a half-finished product, thin documentation, and a deadline that has already slipped.
A takeover is a different job from a new build. Most of the work in the first weeks goes into establishing what actually exists, what was actually promised, and which of those two the remaining budget can still reach.
How it runs
Read everything first. Map what exists against what was promised. Stabilize the riskiest behavior with automated tests, then finish the remaining scope in documented milestones.
Reading everything means the source code, the repository history, the deployment setup, whatever specification exists, and the threads where scope was agreed along the way. On a stalled project the commit history is usually the most honest record you have, because it shows what was genuinely worked on and the point at which the work stopped.
The first two weeks
The opening phase produces three things: an inventory separating what runs from what only appears to run, a written list of the gaps between the promise and the product, and an estimate for finishing that you can hold us to. If the honest conclusion is that finishing costs more than starting again, we say so in writing, even though that is the smaller engagement.
What changes
Within the first weeks you have written requirements, a tested core, and a plan you can verify progress against. The difference between a rescue and a slow failure is exactly that paperwork.
What is in scope, and what is not
A rescue covers finishing the agreed product, making it safe to run in production, and handing it over in a state you own outright. It does not quietly widen into a redesign. New ideas are welcome, and they are quoted separately, so the finish line stays where it was on the day we started.
Proof
EvitaAmenda is a transport-compliance platform whose build had stalled half-finished. We pieced the requirements back together, completed the system and took it live. It now processes GPS data for real fleets every day.
Read the EvitaAmenda case study
Further reading
What to do this week if your developer disappeared
Rescue or rewrite: how we decide
Taking over a codebase nobody documented
If a project of yours has stalled, bring whatever exists. The first conversation is a diagnosis, and it costs nothing.