SECAI Core · phase 15 of 15 · noun
Rollback
To restore a system to the last known good state.
The Explain card
- Plain English
- A rollback restores a system to the last known good state, undoing a change that broke something or introduced a risk.
- Example
- A newly deployed chatbot version starts leaking fragments of its system prompt after a prompt update. The team rolls back to the previous version within minutes while they investigate.
- Why it matters
- Every AI change is a bet. Fast, tested rollback turns a bad model or a harmful prompt tweak from a crisis into a brief blip, but only if previous versions are kept and the rollback is rehearsed.
- Hook
- The undo button for production: test it before you need it.
Word knowledge
How the word is built, where it came from, and what it sits beside in memory.
In a sentence
After the poisoned model update failed validation, the deployment pipeline initiated a rollback to the previous signed model artifact.
How it is built
- roll base move by turning, later extended to reversing movement or change
- back adverbial base in a reverse direction
Where it came from
- Origin
- an English compound noun formed from the phrasal verb roll back
- Entered the language
- 1930s
- What changed
- It first described reversals of prices or policies and later became a computing term for reversing deployments and other system changes.
How it is spelled
- Pattern
- The noun is written as the closed compound rollback, while the corresponding verb is written as two words, roll back.
Spelled like
- backup
- shutdown
- login
Broken into chunks
-
roll
- rolling
- reroll
-
back
- backup
- fallback
What it sits beside
Same subject
- deployment
- checkpoint
- versioning
- model validation
- incident response
Same shape
- backup
- fallback
- shutdown
closed compound nouns
- rollback
- backup
- shutdown
deployment operations
- rollback
- release
- redeployment
Where it sits in the deck
Phase 15: MLOps, Continuous Delivery, and Operational Resilience
Close the loop — operational practices, drift management, and delivery discipline that keep deployed AI systems healthy, current, and continuously improving.
