## Use Cases and Examples

Most data tools make you work around them — exporting CSVs, writing queries, building reports manually. Databox MCP flips that dynamic: you describe what you need in plain language, and the AI handles the rest. Here are some examples of what that looks like in practice.

### Conversational Data Analysis

Instead of building a report or writing a query, just ask. No SQL, no dashboards, no waiting for an analyst.

> *"Which product category has the highest refund rate, and is it getting worse?"*


The AI queries the `All_Orders` dataset, groups by category and time period, and returns: **Highest Refund Rate: Electronics (12%)**, with an insight that *"the rate increased from 8% in Q1 to 12% in Q3, indicating a worsening trend."*

### Multi-Source Data Merging

Combining data from different sources usually means exporting files, opening a spreadsheet, and doing the join manually. With Databox MCP, you just hand over the files and describe what you want.

> *"Merge these by Date and tell me my daily ROAS."* (given `fb_costs.csv` and `shopify_revenue.csv`)


The AI ingests both files into Databox, joins on the Date field, and calculates the **Return on Ad Spend (Revenue/Cost)** for each day. The insight returned might then be *"Average ROAS was 2.4x, with a critical alert on Oct 12th (ROAS dropped to 0.8x when spend doubled but revenue stayed flat)."*

### Automated Decision Triggers

Databox MCP isn't just for chat — it works with any MCP-compatible automation tool. That means you can build workflows that query your data and act on it, running around the clock without human intervention.

> *"Is the 3-day moving average of ROAS below 1.5?"* — asked every day at 6 AM via n8n


If yes, the workflow pauses the Google Ads campaign and posts a Slack alert to the team — no manual monitoring required.

### Data Quality and Cleanup

Real-world data is messy. Databox MCP can normalize and structure it before ingestion, turning a raw export into something you can actually analyze.

> *"Clean this up and add it to my Q3 Expenses dataset."*


The AI normalizes the data, ingests it into Databox, and surfaces that *"**Software Subscriptions account for 40% of Q3 spend**, significantly higher than Q2."*

### Automated Reporting

Keeping stakeholders informed shouldn't mean someone manually pulling numbers every morning. Set it up once and let it run.

> *"Yesterday's metrics"* — triggered every day at 9 AM via n8n


The result is formatted into a summary and posted to a Slack channel, so your team stays up to date without logging into a dashboard.

These are just a few scenarios — essentially, **any situation where you need to get data into Databox or out of Databox in an intelligent, automated way is a fit for MCP**. Whether it's interactive Q&A with your business data, or using Databox as a decision engine inside an AI agent workflow, MCP provides the bridge to do so.