The Core
A guide to Glozo's search methodology: collaborate with Glozo Agent to iteratively refine requirements and surface the most qualified talent.
Search Methodology. Introduction
Finding the perfect candidate is rarely done in a single click. The Core of Glozo is a conversational search engine that combines semantic analysis with an interactive Glozo Agent. It helps you evolve a vague idea into a precise list of top-tier talent.
How It Works: The Iterative Loop
1. The Initial Prompt (Start Broad)
Start by navigating to New Search. Instead of selecting complex filters immediately, simply enter a general description of the position in the Job Description box.
2. AI Chat & Refinement
After clicking search, you will get a broad list of candidates. At this stage, Match Scores (e.g., 20% or 16%) might be relatively low because the system is matching broadly based on your initial prompt.
On the results page, you will see a Glozo Chat panel on your right. This is your command center for refining the search.
- The Goal: The AI analyzes your initial prompt and identifies missing critical details. It might ask: "Does that sound good for what you're thinking?" or ask for clarification on specific skills like AWS or Docker.
- Your Action: You can select suggested quick replies (e.g., "Include Kubernetes as a requirement") or type your own requirements into the chat.
3. The Query Quality Indicator
Inside the Glozo chat, pay attention to the gradient progress bar labeled "Quality of the request".
- Red: The query is too vague (low precision).
- Green: The query is highly detailed (high precision).
Objective: As you answer the chat's questions, the bar fills up from red to green. The "greener" the bar, the more accurate your search results will be.
4. Updating Candidates
When you provide new details, the chat will process this context.
- Clicking "Show candidates": Once you've refined the criteria, click the blue Show candidates button in the chat.
5. The Result
The candidate list refreshes. You will likely see different people than before, but this time with significantly higher Match Scores (e.g., jumping from 20% to 46%+).
Understanding Match Score
The Match Score is a dynamic metric visible on every candidate card (the circle percentage icon). It evolves with your query quality.
- Low Score (20-50%): Usually means your query is generic. The candidate matches "DevOps," but we don't know if they match your specific needs yet.
- High Score (80%+): Means the candidate hits all the specific data points you defined in the chat (Skills + Location + Experience + Context).
Summary: The Perfect Workflow
- Input: Type a basic request or Job Description.
- Analyze: Look at the initial results and the Quality Bar (likely Red/Yellow).
- Refine: Use the Chat to add missing details until the bar turns Green.
- Update: Click "Show candidates" to generate the final high-scoring list.
- Save: Once satisfied, save the results as a Project.