Skip to main content
AI Assistant Prompt Guide
Teddy Fischer avatar
Written by Teddy Fischer
Updated over a year ago

The LaunchNotes AI Assistant boosts productivity and speeds up time-to-publish on Announcements and Roadmap items. This guide covers recommended workflows, prompts, and content.

Recommended workflows:

The AI Assistant can be most helpful when you need to create an Announcement or Roadmap item from internal unpolished documentation, and when you need to improve the writing of the draft content that has already been written.

  1. Writing a new announcement - start with a prompt and product documentation, and get a draft of the announcement outlined in no time.

  2. Writing a new roadmap item - pair a prompt with notes, research findings, or call recordings to form the foundations of a product 1-pager as your roadmap item.

  3. Improve spelling and grammar - select the content of your announcement and open the AI Assistant. Add a prompt at the beginning to improve the spelling and grammar, and voila!

  4. Improve style - select the content of your announcement or roadmap item and add a prompt to improve the style of the writing for your desired tone and audience.

Recommended Prompts:

Prompts are a key part of how the AI Assistant works to transform documentation into marketing-grade content.

We have found luck with several of the prompts featured in this section, but encourage you to customize and tailor the prompts for your specific workflows, source material, and desired output.

  • Summarize these product changes and highlight the core user benefits in a post: (paste list of changes here)

  • Summarize this product brief by writing an announcement that highlights the changes introduced through the features, the benefits to users, and where to find the feature: (paste product brief here)

  • Change this technical writing into a descriptive summary at an eighth-grade reading level: (paste technical writing here)

    • Technical writing examples include Git descriptions, Jira/Linear tickets, or engineering tasks

Defining subject matter, tone, audience-type, and other meta-attributes of your desired output will improve the quality significantly.

Recommended content:

Source material tends to be abundant - and we've seen fantastic results with everything from a SQL query that represents a new chart in the product, to a PRD that outlines the user problems and specifications.

Sources for content include:

  • Product demo transcripts

  • Git descriptions and commit messages

  • Jira tickets

  • Product requirements documents (PRDs)

  • OKRs

  • Raw code snippets

  • Internal wiki documentation

Did this answer your question?