batteriesincluded.com · Questions & Answers

How do we make sure frontline employees have a real say in an AI tool built for their work?

Bring frontline employees into the design before the workflow is built, and give them a genuine way to challenge it. People doing the work can show an embedded engineer where exceptions arise, what information is missing, and which parts of the job rely on conversation or judgment.

Start with a small group representing different roles or working conditions. Ask them to walk through a recent task using examples they are permitted to share. Then have them try an early version and explain what feels helpful, confusing, or inappropriate. Pay particular attention to moments when the tool assumes a routine answer but the employee knows the situation needs a conversation.

Make the feedback loop visible. Record the concerns, identify who will decide what changes, and report back on what was changed or why it was not. Employees should be able to raise a concern without their manager treating disagreement with the tool as a performance issue.

A Lovable or Replit prototype can make ideas concrete enough to test, but a working screen is not proof that the workflow fits the people using it. The EOS focus on roles, team dynamics, and individual contributions offers a useful standard: the tool should support people in carrying out clear responsibilities, not obscure who is accountable or silence their experience.

Category: AI that stays human

← All questions