JevMade field notes / Decision-design guide

Helping an AI assistant understand vague requests before it acts

This guide explains how to use saved instructions to help an AI assistant figure out missing details. The software checks existing files before asking a person for help or making permanent changes.

Original by angel291592Classification

Bookmark Keep this in your collection.

Leave a noteWhat would you try with this?

Only you can see your notes.

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Independent ElevenLabs affiliate. JevMade may earn a commission.

Credits and license

“Intent-Router” by angel291592. Read the original source.

This expanded guide is an AI-narrated adaptation of the source’s essential explanation, examples and caveats, not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

License: MIT

MIT License

Copyright (c) 2026 angel291592

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Our summary

The author shares saved instructions that help an AI assistant understand exactly what a person wants. Instead of guessing or asking dozens of questions about a vague task, the software checks existing project files or customer records to find the missing details itself.

The instructions sort the request into four clear steps. The assistant searches available records for facts, asks a human only when a choice cannot be undone, and stops completely if a system breaks. Finally, it saves a clear summary file that other tools can read.

This is a design lesson for people organizing AI tasks, not a finished program. The current version relies on basic text instructions. Advanced connections to tools like Jev, an AI that chooses from options rather than writing answers, are only planned for the future.

Key takeaways

Guide the software to check available records before it asks a person for missing information.
Keep a clear record of why the AI assistant moved a task from one step to another.
Test the software with graded examples before trusting its estimates about how certain it is.

More guides →See the experiments →