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

# Evidence

> ClosedLoop AI organizes raw feedback into three evidence types: product insights, strategic intelligence, outcomes. Each answers a different question.

ClosedLoop AI organizes customer evidence into three categories, each answering a different question:

| Type         | What it answers                  | Source                                                                                                          |
| ------------ | -------------------------------- | --------------------------------------------------------------------------------------------------------------- |
| **Insights** | "What are customers asking for?" | Product feedback extracted from conversations: bugs, feature requests, pain points, improvements                |
| **Context**  | "Why are they asking?"           | Strategic intelligence: satisfaction, churn indicators, competitor mentions, buying behavior, decision criteria |
| **Behavior** | "What are they actually doing?"  | Product usage data: feature adoption, engagement patterns, drop-off points                                      |

Together, these three types give you the full picture: what customers say, why they say it, and whether their actions match their words.

## Insights

An insight is a single piece of product feedback extracted from a customer conversation, survey, or support ticket. Each insight is one specific thing a customer said, traced to a verbatim quote, linked to their company in your CRM, and scored by business impact.

**What makes insights useful:**

* **Verbatim quotes**: the exact words the customer used, not a summary or interpretation
* **Business impact**: each insight is flagged if it blocks a deal or threatens retention, with the affected revenue visible
* **Pain + workaround + use case**: not just "what's wrong" but "what they do instead" and "what they're trying to achieve"
* **Auto-classified**: severity, emotion, and product area are assigned by the AI, not manually tagged

A typical 30-minute sales call produces 5-15 insights. Multi-topic conversations produce one insight per topic: each is atomic and independently searchable.

## Context (Strategic Intelligence)

Context captures the business dynamics behind the feedback, things like:

* **Satisfaction**: how happy or frustrated a customer is overall
* **Churn indicators**: why a customer is considering leaving, with urgency level (immediate, considering, evaluating)
* **Competitor mentions**: which competitors come up, what they do better or worse
* **Decision criteria**: what drives the buying decision
* **Win/loss reasons**: why customers chose you or didn't

There are 20 intelligence types covering the full customer journey. Each traces back to a specific conversation and quote.

## Behavior (Product Usage)

Behavior data shows what customers actually do in your product: feature adoption, session patterns, drop-off points. This is the "say vs. do" layer: customers may say they love a feature but never use it, or complain about something they use daily.

<Note>Behavior data requires connecting a product analytics platform (Amplitude, Mixpanel, or similar). Without it, you still have insights and context: behavior adds the third dimension.</Note>

## How they work together

The real power is in the combination. An opportunity that clusters 50 insights from 30 customers, linked to \$1.2M in deal value, with 3 churn indicators and accelerating velocity: that's not a feature request, that's a business case. ClosedLoop AI builds this picture automatically from all three evidence types.