> ## Documentation Index
> Fetch the complete documentation index at: https://docs.withlope.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Understanding match scores and explainability

> How AI Scout scores candidates, what the colors mean, and how every result is explained.

When you run an AI Scout search, every candidate receives a **relevance score** and a per-criteria breakdown. Lope uses color coding so you can instantly see how well a candidate matches — and every score is linked to a natural-language explanation so you understand *why*.

***

## Overall relevance score

The overall score is a percentage (0–100%) representing how well the candidate matches across all the criteria you searched for. It is the average of the individual criterion scores, weighted equally.

### Color coding (overall score)

| Score range   | Color  | Label           |
| ------------- | ------ | --------------- |
| **85%+**      | Green  | Excellent match |
| **70–84%**    | Yellow | Good match      |
| **60–69%**    | Orange | Partial match   |
| **Below 60%** | Red    | Weak match      |

The color appears on the candidate card, the score pill, and the avatar ring — giving you instant visual feedback.

***

## Search criteria

AI Scout evaluates candidates on up to five criteria, depending on what you searched for. Only criteria you actually include in your search produce scores and summary sentences.

| Criterion               | What it measures                                                            | Source data                     |
| ----------------------- | --------------------------------------------------------------------------- | ------------------------------- |
| **Job title**           | How closely the candidate's current role matches the title you searched for | LinkedIn experience             |
| **Skills**              | How many of the required skills the candidate has                           | LinkedIn skills, experience, CV |
| **Location**            | How close the candidate's location is to your target                        | LinkedIn profile location       |
| **Years of experience** | Whether the candidate's total experience meets your requirement             | LinkedIn experience history     |
| **Industry**            | Whether the candidate's current company operates in the target industry     | Company data                    |

***

## Per-criterion scoring and colors

Each criterion has its own thresholds. The per-criterion colors appear as **highlighted chips** in the candidate summary and as colored cells in the table view.

### Job title

| Score         | Color  | Meaning                                     |
| ------------- | ------ | ------------------------------------------- |
| **75%+**      | Green  | Strong match — current role closely aligns  |
| **50–74%**    | Orange | Partial — role is related but not identical |
| **Below 50%** | Red    | Weak — current role differs significantly   |

### Skills

| Match rate                          | Color  | Meaning                                           |
| ----------------------------------- | ------ | ------------------------------------------------- |
| **70%+** of required skills matched | Green  | Strong — has most or all required skills          |
| **40–69%** matched                  | Orange | Partial — some key skills present, others missing |
| **Below 40%** matched               | Red    | Weak — few or none of the required skills         |

### Location

| Score         | Color  | Meaning                                           |
| ------------- | ------ | ------------------------------------------------- |
| **50%+**      | Green  | Matches — candidate is in or near target location |
| **30–49%**    | Orange | Nearby — within the broader region                |
| **Below 30%** | Red    | Distant — outside target location                 |

### Years of experience

| Score         | Color  | Meaning                               |
| ------------- | ------ | ------------------------------------- |
| **80%+**      | Green  | Meets the experience requirement      |
| **50–79%**    | Orange | Close — slightly under or over        |
| **Below 50%** | Red    | Gap — significantly below requirement |

### Industry

| Score         | Color  | Meaning                      |
| ------------- | ------ | ---------------------------- |
| **50%+**      | Green  | Works in the target industry |
| **30–49%**    | Orange | Related industry             |
| **Below 30%** | Red    | Different industry           |

***

## How the summary works

Each candidate card includes a natural-language summary that explains the match. The summary is **fully explainable** — every claim is grounded in the score data.

### Structure

1. **Opening line** — Overall relevance percentage and number of criteria evaluated. *Example: "Excellent match across 4 criteria (87%)."*
2. **Per-criterion sentences** — One sentence per active criterion, ordered by score (strengths first). Each sentence names the specific data point and explains the match. *Example: "Currently a Senior Backend Engineer — strong match with your search."*
3. **Highlighted chips** — Key data points inside the summary are highlighted with color-coded chips. The chip color matches the criterion's score tier (green, orange, or red).

### Example summary

> **Excellent match across 4 criteria (87%).** Currently a **Senior Backend Engineer** — strong match with your search. Has **all 4 required skills: Python, Django, AWS, and PostgreSQL**. Based in **Berlin** — matches your location criteria. Has **6 years of experience**, meeting your 5+ years requirement.

In the summary above:

* "Senior Backend Engineer" would be a **green chip** (job title match)
* "all 4 required skills..." would be a **green chip** (skills match)
* "Berlin" would be a **green chip** (location match)
* "6 years of experience" would be a **green chip** (experience match)

If a criterion scored lower, the chip would be **orange** or **red**, and the sentence would explain the gap.

### Polished summaries

After the initial deterministic summary is generated, Lope sends it to an LLM for natural-language polishing. The polished version reads more naturally while preserving all the same data points and color-coded chips. The polishing happens in the background — you see the template summary first, then it's replaced with the polished version.

***

## Score contribution

Each criterion contributes equally to the overall score. The contribution is calculated as:

```text theme={null}
contribution = criterion_score × (100 / number_of_active_criteria)
```

For example, with 5 criteria, each contributes up to 20% of the overall score. With 3 criteria, each contributes up to \~33%.

Hover over a highlighted chip in the summary to see the tooltip showing the criterion's score and its contribution to the overall percentage.

***

## Source links

Each highlighted chip in the summary links back to the source data. Hover over a chip to see:

* **Job title / Years** — Links to the candidate's LinkedIn experience section
* **Skills** — Links to the LinkedIn skills section
* **Location** — Links to the candidate's LinkedIn profile
* **Industry** — Links to the company's LinkedIn page and website

This ensures every claim in the summary can be verified against the original source.

***

## Table view vs. list view

AI Scout results can be viewed in two layouts:

* **List view (cards)** — Shows the full summary with color-coded chips and avatar ring
* **Table view** — Shows individual score columns (Job title, Skills, Location, Experience, Industry) with colored pills

Both views use the same scoring system and color thresholds.