How to Create a Dashboard in Tableau with AI
Creating a dashboard in Tableau is a great way to visualize your business data, but starting with a blank canvas can feel overwhelming. Fortunately, Tableau's built-in AI tools can help you build dashboards faster, uncover patterns you might have missed, and get answers from your data without writing a single calculation. This article will walk you through exactly how to use Tableau's AI to build an effective dashboard, step by step.
First, What Exactly Is "AI" in Tableau?
Tableau doesn't have a single "AI button." Instead, AI is woven into a set of features designed to make data analysis more accessible, work a bit more like a conversation, and accelerate the process of finding insights. For a dashboard builder, the features you'll most likely interact with are part of the Tableau Einstein family.
Think of it in these simple terms:
- Einstein Copilot: This is your conversational AI assistant. It’s a chat pane right inside Tableau where you can ask questions in plain English - like "what were my total sales last quarter?" - and it will generate a chart for you. It's the primary tool for rapidly creating visualizations.
- Explain Data: When you see something interesting in your data - like an unexpected spike in sales - you can select that data point, and this feature will use AI to automatically look for potential explanations and drivers in your other data fields.
- Tableau Pulse: This feature uses AI to automatically surface insights and deliver personalized data digests to business users in a simple, readable format. While you don't build with it directly in the dashboard view, it's powered by the underlying data and AI models.
Essentially, Tableau's AI aims to be your data analysis partner, helping you skip the tedious parts of chart configuration so you can focus on what the data actually means.
Before You Begin: Prepping Your Data for AI
An AI is only as smart as the data it’s given. Before you start asking questions, a little bit of prep work will make your results dramatically better. This isn't about complex data engineering, it's about making your data understandable.
Keep Your Naming Conventions Clear
The AI relies on column headers to understand what you're asking. If you ask, "Show me revenue by customer segment," it’s looking for columns named something like "Revenue" and "Customer Segment."
- Good Naming:
Order Date,Sales Revenue,Customer City,Ad Campaign Name - Confusing Naming:
ord_dt,s_rev_1,cust_geo_loc,cmp_str
Descriptive, human-readable names make it far easier for the AI to correctly interpret your requests.
Check Your Data Types
Ensure your data is properly formatted in Tableau's Data Source pane. Tableau is usually good at guessing, but you should double-check:
- Dates should be recognized as a date or date/time field.
- Geographical data (like City, State, Country) should have the globe icon.
- Numbers you want to measure (like Revenue, Sessions, or Quantity) should be numeric fields.
Getting this right allows the AI to automatically create maps when you ask about states or build time-series charts when you ask about performance "over the last 90 days."
Step-by-Step: Creating a Dashboard with Einstein Copilot
Let's walk through an example. Imagine we've connected a sales data source from a fictional e-commerce store. Our goal is to build a simple sales performance dashboard showing revenue, product performance, and customer location.
Step 1: Connect to Your Data and Open a New Workbook
We'll start this process just like any other Tableau project. Open Tableau, connect to your data source (whether it's an Excel file, a Google Sheet, or a database), and ensure your data loads correctly in the Data Source page.
Once you’re in a new worksheet, you should see the Einstein Copilot icon (a stylized spark/star) in the right-hand panel. Click it to open the chat interface.
Step 2: Ask Your First Question
Don't overthink it. Start with a simple, high-level question. Your prompts don't have to be perfect, you can always refine them later.
In the Einstein Copilot chat box, type a request like:
Show total sales over time
Einstein will analyze your data, identify the fields named "Sales" and a date field (like "Order Date"), and generate a line chart showing sales trends. The chart will appear directly in your worksheet, and the fields it used will automatically populate the Columns and Rows shelves.
Step 3: Refine and Iterate with Follow-Up Questions
Now, let's make that initial chart more specific. Instead of starting over, you build on the previous prompt. Try one of these follow-up commands:
- To change the time frame: "show this by month" or "just for 2023"
- To change the chart type: "change to a bar chart"
- To segment the data: "break it down by product category"
Let’s say we ask Einstein to break it down by product category. The AI will modify the existing line chart to show multiple colored lines - one for each product category (e.g., "Electronics," "Apparel," "Home Goods"). You just turned a simple time-series chart into a segmented trend analysis without touching the drag-and-drop interface.
Step 4: Create a New Worksheet and Ask a Different Question
A good dashboard has multiple visualizations. Let's add an analysis of our top-performing products. Create a new worksheet by clicking the tab at the bottom.
In the new, blank worksheet, ask a new question in the Einstein Copilot:
What are my top 10 products by profit?
Einstein Copilot will create a horizontal bar chart listing the top 10 products, ordered by their profit contribution. Name this worksheet "Top Products by Profit."
Step 5: Let's Build a Map
Geographical data is perfect for a map visualization. Create another new worksheet and ask:
Show me sales by state
Assuming you set up your geographic data types correctly, Einstein should immediately generate a choropleth map (a filled map), where states with higher sales are shaded in a darker color.
Step 6: Assemble Your Dashboard
Now that you have three AI-generated worksheets ("Sales over Time," "Top Products by Profit," and "Sales by State"), it's time to combine them into a dashboard.
- Click the "New Dashboard" icon at the bottom of the screen (it looks like a grid).
- You'll see a blank canvas. On the left side, under "Sheets," you'll see your three worksheets.
- Drag and drop each worksheet onto the canvas. You can arrange them as you see fit - perhaps the sales trend line at the top, with the product bar chart and map below.
You’ve just used natural language to build the components of a dashboard, which you then assembled manually for the final layout.
Step 7: Make It Interactive
Finally, click on the map visualization within your dashboard and select the "Use as Filter" funnel icon that appears. Now, when you click on a state (like California), the "Sales over Time" and "Top Products by Profit" charts will automatically filter to show data only for California. This basic interactivity is a hallmark of a useful dashboard.
Going Beyond Chart Building: Using AI for Deeper Analysis
Building charts quickly is invaluable, but Tableau's AI can do more. Once your dashboard is built, you can use it to ask follow-up questions that lead to real insights.
Use "Explain Data" on an Outlier
Let's say in your "Sales over Time" chart, you see a huge, unexpected spike in sales for the "Electronics" category in October. You want to know why.
Simply right-click on that data point (the peak for October electronics sales) and select Explain Data. The AI will instantly analyze all the other fields in your data set to find likely explanations. It might generate findings like:
- "Sales of 'Smart Phones' were unusually high during this period."
- "A high number of orders came from the 'Corporate' customer segment, which typically doesn't buy electronics."
This transforms your dashboard from a static report into a launchpad for investigation, helping you understand the "why" behind the "what," all without manually building dozens of exploratory charts.
Final Thoughts
Building a dashboard with AI in Tableau transforms the process from a technical chore into a more intuitive conversation with your data. By using Einstein Copilot to ask questions and "Explain Data" to investigate interesting trends, you can assemble powerful, interactive visualizations far more quickly than with the traditional manual approach.
While tools like Tableau offer powerful AI features, we know that the biggest bottleneck is often getting all your data - from Google Analytics, Facebook Ads, Shopify, and your CRM - connected and cleaned in the first place. We built Graphed to solve this by seamlessly unifying your marketing and sales data sources automatically. Once connected, you can ask questions in simple language to create entire dashboards in seconds, not just individual charts, letting our AI do the hard work of analysis, visualization, and insight generation for you.
Related Articles
What SEO Tools Work with Google Analytics?
Discover which SEO tools integrate seamlessly with Google Analytics to provide a comprehensive view of your site's performance. Optimize your SEO strategy now!
Looker Studio vs Metabase: Which BI Tool Actually Fits Your Team?
Looker Studio and Metabase both help you turn raw data into dashboards, but they take completely different approaches. This guide breaks down where each tool fits, what they are good at, and which one matches your actual workflow.
How to Create a Photo Album in Meta Business Suite
How to create a photo album in Meta Business Suite — step-by-step guide to organizing Facebook and Instagram photos into albums for your business page.