Business teams can now put together apps quickly, but they may not know how to secure, deploy, or maintain them. How do you handle ownership when engineering is asked to support one? Do these apps go into your production environment or somewhere separate?
Who should own and host AI-generated apps built by business teams?
Asked by
b5+fuse_912
Answers
Give them a path to deploy, but don't make engineering the automatic owner. The app should go through the same pipeline gates as any other app, including security scans. The team asking to deploy it is responsible for getting failures fixed.
That still leaves a practical question: if the business team can't fix the code, someone has to decide whether engineering has capacity to take it on. Passing a prototype over the wall shouldn't quietly become a support agreement.
Reply to this answer
Keep these as prototypes until there's a case for production. No real data or production access at that stage. A useful prototype can help explain the requirements, then engineering can decide whether to adopt, refactor, or rebuild it.
Reply to this answer
Another option is self-service hosting in separate cloud accounts with guardrails. Let the teams run their own infrastructure, but put security controls and alerts around it.
Data classification matters here. Keep highly sensitive data out of those environments rather than assuming isolation alone makes every app safe. Separate hosting also doesn't eliminate the need for someone to respond when an alert fires.
Reply to this answer
Treat adoption as a funding decision, not just a technical handoff. Look at the business value alongside the engineering effort and ongoing operations budget. A promising app might be worth a heavy refactor or a rewrite, but that work needs a real owner and room in the plan.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Trelwick.2:
Separate accounts sound useful for experimentation, but I'd want the boundaries and response responsibilities agreed before anyone deploys.