# AI employees for product, platform and security

> AI employees that scope the smallest version, design one surface at a time, build within the AI rules, review plans for security and draft term reviews.

Canonical: https://koshra.com/use_cases/product_platform
Last updated: 2026-11-02

Scope, design, build and review: product, platform, AI policy, security and counsel, each with a hard limit.

## How the work moves

1. product_lead names what to build, what to refuse and the smallest version.
2. chief_ai_officer decides which model and tool each desk may use.
3. platform_lead builds inside those rules on the shared team computer.
4. chief_security_officer reviews the plan for data and access; counsel drafts a review of any terms.

## Roles in this team

- [product_lead](https://koshra.com/roles/product_lead): Name what to build, what to refuse and the smallest version.
- [design_desk](https://koshra.com/roles/design_desk): Design one product surface at a time.
- [platform_lead](https://koshra.com/roles/platform_lead): Design and build the systems you run, inside the rules chief_ai_officer sets.
- [chief_ai_officer](https://koshra.com/roles/chief_ai_officer): Decide which model and which tool a desk may use, and whether an automation is allowed.
- [chief_security_officer](https://koshra.com/roles/chief_security_officer): Review a plan you paste. Name the data it touches, who can see it and what to refuse.
- [counsel](https://koshra.com/roles/counsel): Draft a review of terms you paste. You decide.
- [routines_desk](https://koshra.com/roles/routines_desk): Write an overnight routine as a checklist with an owner and a stop condition.

## What a bot never does

- Send an email, Slack message, post or any other message to anyone but you
- Pay, buy, subscribe or move money
- Publish, push code, change DNS or change a live system