What is an AI executive assistant?

An AI executive assistant is a general model plus persistent context (your role, rules, and standards) and reusable prompt templates for recurring work like meeting prep, status updates, and email drafts. It is not autonomous. You stay accountable for output.

What should a system prompt include?

Include your role, what you work on, voice and format preferences, hard rules (privacy, verification), and how you want answers structured. Keep it to one page so you can paste it into ChatGPT projects, Claude projects, or IDE rules.

What is a good task prompt structure?

Use six blocks in order: Role, Context, Task, Constraints, Output format, Quality bar. Role sets who the model should be for this request. Context is minimum background. Task is the single ask. Constraints say what to avoid. Output format defines length and shape. Quality bar defines success.

Home › Tools › AI Assistant Prompt Builder

AI Assistant Prompt Builder

I use this shape every day: a one-page system prompt for who I am, plus a six-block task template for each request. Fill in the fields, copy the output into ChatGPT, Claude, or your IDE, and you have a working executive assistant setup in minutes.

Your profile

Role / title

Work context (company, scope, stakeholders)

Recurring tasks you want help with

Voice and tone

Hard rules

Starter task template

Generated prompt

How to set up your AI executive assistant

Copy the system prompt into a ChatGPT project, Claude project, or rules file in your IDE.
When you have real work, open the task template, paste your notes where indicated, and send it.
When output misses, fix the prompt (usually Context or Output format), not the model brand.
After a week, save your best variants. That library becomes your personal playbook.

For the full walkthrough, read Setting Up an AI Executive Assistant. For enterprise rollout, see the AI program management hub.

The six-block task shape

Role who the model should be for this request
Context minimum background, paste notes here
Task the single outcome you need
Constraints what to avoid or never assume
Output format length, bullets, table, or draft
Quality bar how you will judge success

Built by Arsenii Samoilov, a Senior Technical Program Manager who uses AI daily for program work. Get in touch for Senior TPM roles or AI program advisory.

Read the setup guide →