# AI Analytics

The **AI Analytics** section in BotPenguin helps you understand how effectively your AI Agent is responding to users. It provides insights into user feedback, common conversation topics, and the overall quality of AI responses, enabling continuous improvement of your AI Agent.

{% embed url="<https://youtu.be/k-uJU0luVRk>" %}

***

### 📍 Accessing AI Analytics

1. Log in to your **BotPenguin Dashboard**
2. Go to **Analytics** from the left menu
3. Click on the **AI Analytics** tab

***

### 📊 AI Analytics Sections Explained

#### 1. 👍👎 Thumbs Up and Down

This chart shows direct user feedback on AI responses.

**What it represents:**

* **Thumbs Up** – Users found the AI response helpful and relevant
* **Thumbs Down** – Users were not satisfied with the AI response

**How to use it:**

* Track user satisfaction trends over time
* Identify days or periods with higher negative feedback
* Use this insight to improve responses, intents, or training data

**Filters available:**

* Date range
* Bot selection
* View type (Daily / Weekly / Monthly)

<figure><img src="https://1745791824-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAPDb8cKQtGlIAfgHjcsQ%2Fuploads%2FfymipVTHEkKFY5hENFBz%2FChart1.png?alt=media&#x26;token=274d2650-30b3-41f2-937f-455d9fc9fd11" alt=""><figcaption></figcaption></figure>

***

#### 2. 🧠 Frequent Topics

The **Frequent Topics** section displays a word cloud of the most common words and phrases users mention while chatting with the AI.

**What it represents:**

* Larger words indicate topics discussed more frequently
* Helps identify user intent, interests, and common questions

**Why it matters:**

* Discover what users ask most often
* Identify missing FAQs or knowledge gaps
* Improve AI training content and intent coverage

<figure><img src="https://1745791824-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAPDb8cKQtGlIAfgHjcsQ%2Fuploads%2FIClTu2KzRCziid8abhSl%2FChart2.png?alt=media&#x26;token=3b86473e-2431-4dbd-abff-c3036fe9a4f3" alt=""><figcaption></figcaption></figure>

***

#### 3. 📈 Response Effectiveness

This chart evaluates how well the AI Agent understands and responds to user queries.

**Response categories:**

* **Answered** – Queries correctly understood and answered by the AI
* **Abusive** – Messages flagged as abusive or inappropriate
* **Out of Context** – Queries unrelated to trained knowledge
* **Lack of Context** – Queries with insufficient information for the AI to respond accurately

**What it helps you measure:**

* Overall AI response quality
* Gaps in training data
* Areas where better context handling is needed

<figure><img src="https://1745791824-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAPDb8cKQtGlIAfgHjcsQ%2Fuploads%2FPCJdHr0zNx5a3DYYx239%2FChart3.png?alt=media&#x26;token=158e5a1c-f7be-49bc-8ab2-e9d76843ec26" alt=""><figcaption></figcaption></figure>

***

### 🎯 Use Cases

* Improve AI training accuracy
* Reduce incorrect or irrelevant responses
* Identify popular user queries
* Measure AI Agent performance and reliability
* Enhance overall user satisfaction