evenwing

You are a market research analyst covering technology.

Task

1. Infer the closest software or technology market category from the user input, then list the top 3 players.
2. Do not ask clarifying questions. Respond with a brief apology if the input is empty or entirely unrelated to technology.
3. Your ranking priority based on information available: revenue -> valuation -> number of customers -> number of G2/app store ratings

Output Requirements

evenwing

Cluster the raw notes from the long-form-annotation field into an organized set of distinct trace codes 

1. Create a taxonomy that can be used to systematically label new traces with a binary (yes/no) judgment. 
2. Include paraphrased snippets from the output field as examples for each category.

Example

Trace Code: Category Match

• Definition: The category chosen is a good match for the user input
• Example: for input “AI-Native customer Support” the ranked companies were startups < 5 years old

evenwing

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

evenwing

Data annotation app

GOAL

Build a web application that allows human reviewers to inspect multi-turn conversational traces and record structured annotations. Focus on fast, minimal review workflow. Avoid unnecessary features.

DATA

An example set of traces is attached.

evenwing

# Sample AI PRD: Support Ticket Triage (v1.0)

**Owner:** [Name]  | **Status:** Prototyping  | **Default Model:** Gemini 3 Flash Preview

## 1. Problem & Business Value

Support leads spend about 4 hours per day manually tagging tickets. This creates a lag before specialists can see high-priority issues.

**Proposed solution:** A background agent that autonomously classifies tickets in real time by Intent, Sentiment and Urgency