911 program managers need a retire-and-replace plan when an old intent model is superseded. Which lifecycle concern does this address?

911 program managers needing a retire-and-replace plan when an old intent model is superseded are addressing model retirement and replacement in the AI lifecycle: decommission the prior production model, cut over traffic to the successor, retain artifacts as required for audit, and avoid dual conflicting authorities answering the same dispatch prompts. That concern is not model retirement and replacement when a successor is trained, which completes learning but does not validate performance or promote the successor to production to supersede the old model. Model retirement and replacement when a successor is staged is only pre-production preparation, not the promotion cutover that decommissions the superseded model and directs live traffic to its successor. Model retirement and replacement when a successor is validated confirms testing is passed, but validation alone does not decommission the prior production model or complete promotion to serve dispatch prompts. Exam caveat: retirement includes more than deleting a file—version pinning, rollback capability, data retention, runbooks, and communication to dispatch stakeholders matter in public-safety settings. Operational check: document successor promotion criteria, run a controlled cutover with rollback, archive the retired intent model per policy, and confirm only the approved model serves 911 intent traffic after replacement.