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

# Signals

> Real changes in how AI talks about you, not day-to-day noise.

AI answers vary from one day to the next. Signals separate real changes from that noise, and explain them.

## Observability

Observability compares the last 14 days with the 14 days before, overall and by platform, topic and persona.

A change only becomes a signal when it's large and backed by enough answers to be real:

| Metric | Flagged when |
| - | - |
| Mention rate and citation rate | The move is at least 5 points, there are at least 20 answers in each window, and the change is statistically significant |
| Average position | The move is at least half a position, with at least 15 mentions in each window, and the change is statistically significant |

Each new signal explains:

* **What happened**: the change, in plain English.
* **Why it happened**: the evidence behind it, such as questions that moved, competitors in the same answers, or cited sites that changed.
* **What to do**: the next step.

Signals are checked every night. Mark them useful, not useful or acted on to keep the feed focused, or dismiss them.

## Brand content

Brand content groups opportunities found in the data Sightline has already collected, each with where it happens and an action:

* Page metadata
* Content relevance
* Competitive presence
* Brand protection on review and social sites
* Search ranking
* Content gaps
* Suggested competitors
* Question coverage

> [!TIP]
>
> Read the "what to do" on any negative signal first.


## Related topics

- [How Sightline works](https://docs.asksightline.com/how-it-works.md)
- [Tracked questions](https://docs.asksightline.com/concepts/questions.md)
- [Quickstart](https://docs.asksightline.com/quickstart.md)
- [MCP connector](https://docs.asksightline.com/integrations/mcp.md)
- [Plans](https://docs.asksightline.com/plans.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.