batteriesincluded.com · Questions & Answers

What should we own and have documented before our forward-deployed AI engineer finishes the engagement?

Make handover part of the engagement agreement, not a final-week cleanup task. Your company should retain control of the application, its operating accounts, and the information another engineer would need to maintain it.

For a tool built with Lovable or Replit, clarify ownership and access for the workspace, source code, database, hosting, domain, integrations, and any AI provider accounts. Check the relevant platform terms rather than assuming that access to a project means you control every dependency.

Require these deliverables:

- A written agreement covering ownership of custom code and any third-party licensing restrictions.
- Company-controlled administrative accounts, with appropriate backup access.
- A deployment guide explaining how changes reach production and how to reverse them.
- A map of integrations, credentials, recurring charges, and data storage locations.
- Instructions for backups, recovery, common failures, and support escalation.
- A known-issues list separating operational risks from optional improvements.

Keep EOS as the operational backbone. Assign ongoing application accountability to an appropriate seat on the Accountability Chart. That person does not need to become a developer, but must know how to get a problem resolved.

Before sign-off, ask someone other than the departing engineer to follow the documentation and complete a routine maintenance task in a safe environment. If they cannot, the handover is incomplete.

The goal is continuity, not a promise that software will never need maintenance. Forward-deployed engineering should leave your team more capable, not dependent on one person's memory.

Category: Forward-Deployed Engineering

← All questions