> For the complete documentation index, see [llms.txt](https://www.analytics-docs.laplacetec.com/guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.analytics-docs.laplacetec.com/guide/en/mcp/use.md).
# Try asking these after connecting
Once you've connected, just ask questions in your usual AI chat window. The AI queries the data organized in Laplace directly and answers. No need to download to Excel and paste.
{% hint style="info" %}
If you haven't connected yet, start with [Getting started with MCP](/guide/en/mcp/start.md).
{% endhint %}
## Check the connection with a first question
First, use the question below to check that the connection works.
{% code overflow="wrap" %}
```
Show me the list of data sources connected to Laplace.
```
{% endcode %}
If you see the stores and ad channels you've connected, you're all set.
{% hint style="info" %}
If you connected with Laplace account login and have several projects, ask "Show me my Laplace project list" and then tell the AI which project to analyze.
{% endhint %}
## What the AI can do
| Category | What it can do |
| ------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Data queries & analysis | Query metrics from data sources such as orders, ads, and GA4 by period, channel, or product; compare periods; cohort, cross-purchase, customer statistics, and flow analysis; analysis combining multiple data sources |
| Custom dashboards | Find and read dashboards; create and edit dashboards from your analysis |
| [Chart notes](/guide/en/customdashboard/custom/memo.md) | List date notes such as promotions and events; add, edit, and delete notes |
| Costs & fees | Query registered product costs and store fee history |
| Product & ad matching | Check option ↔ base product matching and ad matching status; edit matches |
| Segments | Query saved customer segments |
| Data change history | Query data management change history and restore points |
You can see the full tool list and example questions in Laplace **Settings** › **Developer** › **Tools**. Search by what you want to do or by question, and filter by **Read only** / **Can change data**.
{% hint style="warning" %}
Personal information such as buyer names, emails, phone numbers, and addresses is never returned.\
The AI can also change data, for example by creating dashboards or editing matches. For the exact scope, see [Security and permissions](/guide/en/mcp/security.md).
{% endhint %}
## The formula for a good question
The more specific your question, the more accurate the answer. Try including these elements.
{% hint style="info" %}
```
Period + channel / scope + metric + comparison baseline + goal
```
{% endhint %}
| Question | Result |
| --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- |
| ❌ Analyze sales | The AI doesn't know what to look at or for which period, so you only get a basic summary |
| 🔶 Tell me last month's Smartstore sales | You get the numbers, but not why |
| 🟢 Compare last month with the month before, and split products that had ad spend into those with positive contribution margin and those at a loss, by channel. I want to see whether any products have high ROAS but low contribution margin | You get an answer that covers the comparison, causes, and actions |
## Example questions by role
Copy these and just change the period and channels.
### CEO
{% code title="Monthly business summary" overflow="wrap" %}
```
Put last month's sales, ad spend, and ROAS by sales channel into a table compared with the previous month. Tell me first which channel changed the most and why.
```
{% endcode %}
### Marketer
{% code title="Ad campaign check" overflow="wrap" %}
```
Show me ad spend and ROAS by campaign for Meta and Google ads over the last 30 days. Tell me the 5 campaigns whose ROAS dropped the most compared with the previous 30 days, and possible causes.
```
{% endcode %}
### MD / Product manager
{% code title="Product profitability" overflow="wrap" %}
```
Show me the top 10 and bottom 10 products by sales and contribution margin last month. If any products have no cost registered, list them separately.
```
{% endcode %}
### CRM / Retention manager
{% code title="Repurchase cohort" overflow="wrap" %}
```
Show the monthly repurchase rate of customers who first purchased in January this year as a cohort, and compare it with customers who first purchased in December.
```
{% endcode %}
### Save it as a dashboard
{% code title="Turn an analysis into a dashboard" overflow="wrap" %}
```
Make the sales, ad spend, and ROAS comparison by channel we just looked at into a custom dashboard. Name it "Monthly channel check".
```
{% endcode %}
### Leave a chart note
{% code title="Record a promotion period" overflow="wrap" %}
```
Record last month's promotion period as chart notes. Write "Spring promotion starts" on the start date and "Spring promotion ends" on the end date.
```
{% endcode %}
Notes you leave also appear on custom dashboard charts. For which charts show them and how to use the screen, see [Chart notes](/guide/en/customdashboard/custom/memo.md).
For more question templates, you can also use the [role templates in the AI agent guide](/guide/en/ai_agent/basic_guide.md#role-templates) as is.
## Tips for better results
1. **Ask using the terms shown in Laplace**\
Using metric names from Laplace dashboards, such as "contribution margin", "net sales", or "BEP ROAS", helps the AI find things more accurately.
2. **Split long periods**\
We recommend querying 90 days or less at a time. For a 1-year trend, asking quarter by quarter is more reliable.
3. **Dig deeper with follow-up questions**\
After you get an answer, follow up with "why is that?" or "break it down by product" so you don't have to explain everything again.
4. **Tell the AI your role**\
Giving a perspective such as "from a marketer's point of view" or "for a CEO report" tailors the answer format.
## Ask your dedicated manager
Not sure what to ask?
Contact us through the inquiry page below. Your account manager will show you how to make the most of it.
> 💬 [Contact us on our website](https://www.analytics.laplacetec.com/contact)
**Email:** [**product@laplacetec.com**](mailto:product@laplacetec.com)