# Chat

Data is stored in complex data models and insights are scattered across hundreds or more dashboards and visualizations. The few experts that can navigate this complexity become a bottleneck for the organization.

Instead of creating Jira-Tickets or following up 3 times on an e-mail, you can just ask Dot - the data bot. Dot is trained by the data team to answer all the flexbile and nuanced questions about your business you care about.

### What can you ask?

Let's look at some potential questions for the marketing team of an online-shop.

1. **Ask for help on how to analyze data.**

<figure><img src="https://3709841693-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCeRlmJf2y704TpjpubE1%2Fuploads%2FO0ipGIxrt3O1zjrbClbY%2Fimage.png?alt=media&#x26;token=7070b16f-4332-4f81-b78d-60b36a4968a4" alt=""><figcaption></figcaption></figure>

2. **Ask questions that need a deep understanding of your companies data model.**

<figure><img src="https://3709841693-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCeRlmJf2y704TpjpubE1%2Fuploads%2FOldeW5Novytl08Jc6lQ9%2Fimage.png?alt=media&#x26;token=da5ca638-66de-4033-ae7e-9710f9c52bb3" alt=""><figcaption></figcaption></figure>

3. **Request not just specific data, but also specific visualizations**

<figure><img src="https://3709841693-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCeRlmJf2y704TpjpubE1%2Fuploads%2F5CnlruaVtLByXjmooGLy%2Fimage.png?alt=media&#x26;token=fc4d2c9d-64d1-4424-a93c-adf3af936710" alt=""><figcaption></figcaption></figure>

4. **Look up detailed information about customers, orders, ...**

<figure><img src="https://3709841693-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCeRlmJf2y704TpjpubE1%2Fuploads%2FejFbeDUbAz2M1eh9ReZk%2Fimage.png?alt=media&#x26;token=318969d6-cbfc-4813-ade1-6f0863fa4cc7" alt=""><figcaption></figcaption></figure>

5. **Ask for anything. Dot will not hallucinate, because it is grounded by the connected data.**

<figure><img src="https://3709841693-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCeRlmJf2y704TpjpubE1%2Fuploads%2FnFD0mPfXeJkprAYIofwv%2Fimage.png?alt=media&#x26;token=5612371f-a346-48f3-88e0-26a559a1d744" alt=""><figcaption></figcaption></figure>

6. **If Dot didn't find what you were looking for, just say "Try again."**

   Dot is based on current frontier models and statistically in nature. By trying again it might choose another way to approach the question and succeed.

### How should you ask?

**Ask Clear and Precise Questions**

To get the most accurate and relevant answers from Dot, follow these best practices:

* **Be Specific:** Clearly define what youʼre asking for.

  *Example:* Instead of saying "I need data on suppliers," specify "I need a list of active suppliers in Region 1 for the last month."
* **Provide Context:** Include relevant details or constraints.

  *Example:* If you need data for a specific time period, mention it explicitly.
* **Clarify Ambiguities:** If your question could be interpreted in multiple ways, provide additional detail.

  *Example:* If you request "order data," clarify whether you mean GMV, number of orders, or another metric.
* **Use Business Terminology:** Stick to the terms from the companyʼs business glossary.

  *Example:* Use "Active Buyer" instead of "buyer."
* **Ask for Help:** If youʼre unsure how to phrase your request or what data is available, ask Dot for help.

#### 💯 **Ideal Question Components**

To get the most accurate results from Dot, ensure your questions include the following components (if possible):

* **Which metrics?** Specify the exact metric youʼre looking for (e.g., revenue, number of orders, active buyers).
* **What entity?** Specify the entity you're asking about (e.g., suppliers, products, buyers).
* **During which time period?** Specify the time frame (e.g., last month, Q1 2024, July 2024.
* **At which granularity?** Specify the level of detail needed (e.g., daily, weekly, monthly, region-level, country-level).
* **For which location, demographic, or specific entity?** Identify any geographic areas, demographic groups, or specific entities you want to focus on (e.g., Region 1, Paris, CUSTX).
* **Using which unit?** Specify the unit of measurement for the metric (e.g., Euros, Kilograms).

#### ✅ **Examples of Good Questions**

* **What was our supplier revenue in euros for July 2024 per region in region 2 and 3?**
* **How many active premium suppliers did we have in Paris in March 2024?**
* **What is the % active users per region and per month for the first 6 months of 2024?**

#### ❌ Examples of Bad Questions

* **What is our current revenue?**
  * *Reason:* No timeframe provided.
* **Which customer in Region 5 has been using our product the longest?**
  * *Reason:* Ambiguous, does "longest" refer to since sign-up or since the first order? Which customer, buyer or supplier?
* **Can you tell me the best selling product for next month?**
  * *Reason:* Dot does not provide future predictions.
* **Show me all orders ever placed in the system.**
  * *Reason:* Too broad, lacks a specific timeframe or criteria.

#### :question:More help

If you're uncertain about Dot's answer, just ask the data team for help!

They can audit the query and data model to double check the analysis.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.getdot.ai/using-dot/chat.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
