You have more than a working system
After a successful first project you have an evaluation discipline, a monitoring pattern, a team who understands the trade-offs and credibility with whoever approves budget.
That is worth as much as the system itself and it should shape what comes next.
Reuse deliberately
- The same evaluation harness, extended
- The same monitoring and alerting pattern
- The same review interface conventions, so nothing is relearned
- The same audit trail structure
A second project reusing the first's foundations typically costs 40–60% of what it would standalone. That is the real return on doing the first one properly.
Raise the stakes one step
| First project | Sensible second |
|---|---|
| Internal, low consequence | Internal, higher volume or value |
| Assist only | Assist plus narrow automation |
| One document type | Two or three types |
| One system | Two systems connected |
One dimension at a time. Increasing all four at once is how a confident team produces its first failure.
Choose from evidence now
The first project gave you data: where corrections cluster, what people ask for, where the queue backs up. Let that choose the second rather than a fresh brainstorm.
The unanswered-question log and the correction analysis are usually pointing at it already.
Keep the first one healthy
A common failure is moving attention entirely to the new project while the first drifts. Its maintenance is a few hours a month and it should stay booked.
A degraded first project undermines the case for the third, whatever the second achieves.