How to Make a Graph in Looker with AI

You no longer need to be a data expert to build powerful visualizations in Looker. With its new AI-powered features, you can create graphs and charts by describing what you want to see in plain English. This guide will walk you through exactly how to make a graph in Looker with AI, from your first simple question to refining it into a dashboard-ready chart.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Data Analysis in Looker: The Old Way vs. The New Way

For years, creating visualizations in Looker meant using the "Explore" interface. While incredibly powerful, it came with a learning curve. You had to know your way around the data model, understand the difference between dimensions and measures, and manually drag-and-drop fields to build the perfect report. If you didn't know the exact name of a field, you'd be scrolling through long lists to find it.

This process worked well for data analysts but could feel intimidating for marketing managers, sales leaders, or anyone else who just wanted a quick answer to a business question. The barrier wasn't a lack of curiosity, it was a lack of time and technical training.

The introduction of generative AI changes this completely. Now, that same process looks more like a conversation:

• Before AI: Open an Explore, find the "Sales amount" measure in the left-hand panel, find the "Order Date" dimension, add a filter for the last 12 months, select a "Line Chart" visualization, and then run the query.
• With AI: Type "Show me sales over the past year as a line chart" and get the result instantly.

This shift makes data exploration more intuitive, accessible, and much, much faster. It empowers your entire team to be more data-driven by removing the technical hurdles that once stood in the way.

A Step-by-Step Guide to Creating a Graph in Looker with AI

You can create visualizations from natural language using the AI Assistant in Looker's Explore interface. Access to this feature depends on your organization's Google Cloud-specific "Generative AI in Looker" settings, so you may need to check with your administrator if you don't see the option. Once you've confirmed you have access, you're ready to start building.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

1. Open an Explore and Find the AI Assistant

First, navigate to the dataset you want to analyze. From your Looker homepage or a relevant dashboard, select an "Explore." This is the same starting point you would use for manual chart building.

Once the Explore page loads, you will see the familiar layout with your field picker on the left and the visualization area on the right. Look for a prompt box or a button labeled "Explore with AI Assistant," "Ask questions with AI," or something similar. This is your gateway to the natural language feature.

2. Start With a Simple, Clear Question

Think about what you want to know, and just ask. You don't need to know the specific field names or data terminology. The AI is designed to understand general business language and translate it into a formal query.

Here are a few examples of great starting prompts:

• "Show me our total revenue by month for last year"
• "What are our top 10 products by units sold?"
• "How many new users signed up each week in Q3?"
• "Compare traffic from Google versus Facebook Ads over the last 30 days"

Type your question into the AI Assistant's prompt box and hit enter. In the background, Looker translates your plain-English request into LookML code, runs the query, and generates a visualization that best fits the data. For "revenue by month," you'll likely get a line chart. For "top 10 products," you'll probably see a bar chart.

3. Edit and Refine Your Graph Through Conversation

The real power of this feature is that it's iterative. Your first chart is rarely your last. The initial result often sparks new questions or ideas for drilling down deeper. You can refine your graph simply by asking follow-up questions.

Let's walk through an example. Suppose you start with a broad prompt:

Your Prompt: "Show me customer traffic by country"

Looker might generate a table or a map visualization. From here, you can continue the conversation to customize it:

• Your follow-up: "Ok, change this to a vertical bar chart." The assistant will instantly rebuild the visualization as a bar chart.
• Your follow-up: "Only show the top 8 countries." The chart will update to filter for just the highest traffic sources.
• Your follow-up: "Now stack the bars by device type." The assistant will add another dimension, showing you the breakdown of mobile vs. desktop traffic within each country's bar.

This back-and-forth process is incredibly efficient. What once required dozens of clicks—adding filters, adjusting pivots, changing visualization settings—can now be done in seconds with a few simple sentences. You are quite literally just talking to your data.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

4. Make Final Tweaks Manually (If Needed)

The AI gets you about 95% of the way to a finished product. For final polish, you can still use the classic interface. Once the AI has generated the core chart, you can click on the "Edit" button to open the familiar visualization settings panel. From there, you can:

• Adjust colors to match your company's brand guidelines.
• Edit axis labels for clarity.
• Change the scale (e.g., from linear to logarithmic).
• Add trend lines or reference lines.

Once you're happy with the result, you can save your new graph as a Look or add it directly to a dashboard to share with your team.

Tips for Getting the Best Results

While Looker's AI is powerful, you can help it help you. Here are a few tips for getting accurate and efficient results from question to insight.

Be Vague or Be Specific - Both Can Work

You don't need special skills to write the perfect "prompt." You can start with a very simple, six-word question like "What were my sales last month?" The AI assistant is smart enough to find the primary sales metric and create a default visualization. This is a great way to start exploring a dataset you're not familiar with.

However, if you know exactly what you want, you can get there in a single step by providing more context. For instance:

Show me Monthly Recurring Revenue for the "Enterprise" customer tier compared to the "SMB" tier as a line chart covering the past two years.

This detailed prompt provides the metric (MRR), the segments to compare (Enterprise vs. SMB), the chart type (line chart), and the timeframe, allowing the AI to build the perfect chart in one shot.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Use the Generated Code as a Learning Tool

One of the hidden benefits of the AI assistant is its ability to teach you how Looker works. After the AI generates a chart, you can often view the underlying code it created. This shows you exactly which dimensions, measures, and filters were used to answer your question. By observing how your natural language gets translated, you can quickly learn the specific terminology and structure of your company's data model, making your future analyses even more efficient.

Understand Its Limitations

Generative AI in analytics is a huge step forward, but it's not magic. Its effectiveness depends entirely on the quality and structure of the underlying data. The AI assistant can only analyze data that has been properly modeled in Looker. If dimensions and measures are mislabeled, or if the relationships between data tables are not clearly defined by your data team, the AI will struggle to return accurate results.

If you ask a question and get a strange or incorrect result, it's often a sign that there is an issue with the data model itself, not with your prompt.

Final Thoughts

Making a graph in Looker with AI transforms data analysis from a technical task into a simple conversation. You can now explore your data, ask follow-up questions, and build complex visualizations using plain English, massively speeding up the time it takes to get from a question to an insight.

We built Graphed to push this conversational approach to data even further. We believe you shouldn't have to live inside a single ecosystem like Looker or even worry about where your data is stored. Graphed connects to all your tools - from Google Analytics and Shopify to your CRM and ad platforms - in just a few clicks. Our AI is designed from the ground up to let you instantly build real-time, cross-platform dashboards and get answers using natural language, making powerful analysis accessible to everyone on your team, regardless of their technical background.