Data Exploration and AI-Generated Datasets

Simon AI can query your warehouse directly to answer questions about your customer data: purchase patterns, profile completeness, behavioral trends. It can also turn those answers into reusable AI-generated datasets you can use in segments, journeys, and personalization. No SQL or data team involvement required.

This capability is part of AI Chat.

🟦

Core AI · Included with your Simon platform.

Getting started

Navigate to AI Studio > Chats and start a new conversation. Describe what you want to explore or build, and the agent queries your data and surfaces the answer in the chat.

📘

No SQL required

Describe what you want in plain language. The chat writes the queries, analyzes results, and presents answers in the conversation, and can save those insights as AI-generated datasets your whole team can use.

Example prompts

• "What are the top product categories purchased by customers in the last 30 days?"
• "Create a dataset that scores customers on cold-weather readiness"
• "Show me the purchase frequency distribution for loyalty members"
• "What percentage of my contacts have made a purchase in the last 90 days?"
• "Which customers have bought more than three times but not in the past 60 days?"
• "Create a dataset of customers who purchased in the last 30 days with their order details"
• "Show me what datasets I have available"

What you can do

Behavioral analysis

The chat can analyze how your customers behave over time: not just what fields exist in your data, but what patterns your data reveals. Use this to understand your base before building audience strategy.

Example prompts

• "What does the purchase frequency distribution look like for my top 20% of buyers?"
• "Which product categories have the highest repeat purchase rate?"
• "What's the average time between first and second purchase for customers acquired through email?"
• "Show me customers who purchased more than three times but haven't bought in 60 days"

If a behavioral pattern is worth capturing as a reusable signal, you can ask the agent to create an AI-generated dataset from that logic, and it stays available for segmentation and journey targeting going forward.

Contact profile analysis

The chat can query the distribution and completeness of profile attributes across your customer base, useful for understanding data quality, identifying enrichment opportunities, and informing segmentation strategy.

Example prompts

• "What percentage of my contacts have a verified email address?"
• "Show me the distribution of loyalty tier across my active contacts"
• "How complete is my demographic data for customers acquired in the last 12 months?"
• "What's the opt-in rate breakdown by acquisition channel?"

How AI-generated datasets are created

1. Describe the field — or set of fields — you want in natural language
2. The agent searches and analyzes your warehouse tables
3. It outlines the proposed dataset: logic, inputs, and sample values
4. Review and approve the proposal
5. View statistics and distribution charts on each field's details page
6. Use the fields in segments, journeys, and personalization

The result is a composable dataset like any other in your account — same refresh behavior, same dependency tracking, same availability everywhere your data is used. A single request can produce a dataset with multiple related fields.

Working inside a Project

When you open a chat from within a Project, the agent uses your Project's campaign objective as context for data exploration and field creation. Your organization's AI Context is also applied automatically.

Availability

Data exploration and AI-generated dataset creation is available to all AI-enabled organizations.