Update: AI Metric Creation and Filter Search

This update adds two new features.

We have added a feature that automatically generates a metric when you describe the one you want in natural language, and a feature that automatically sets filters and runs a search when you enter search conditions in natural language.

Why this matters

Creating a metric requires manually combining several configuration items such as a name, conditions, and filters. When searching data, you also need to configure filters in detail to reach the results you want.

This update replaces that process with a single line of natural language, focusing on supporting faster and more accurate decision-making.

Automatic metric creation with natural language

Describe the metric you want in words, and the AI analyzes it and creates the metric automatically.

• Natural language input: In the metric creation flow, you can enter the metric you want in natural language, such as "repurchase rate by channel over the last 30 days."
• Automatic metric creation: The AI analyzes your input and automatically configures the metric name, conditions, filters, and other settings.
• Reflected in the ontology: Automatically generated metrics are stored just like existing metrics, and can be viewed and managed the same way afterward.

Now, instead of filling in each configuration item one by one, you simply enter the metric you want and it is created immediately. The time you spent designing metrics can now go toward looking at them and making decisions.

Filter search with natural language

Enter the search conditions you want in natural language, and the AI analyzes them, sets the filters automatically, and runs the search.

• Natural language to filters automatically: Enter conditions such as "unresolved inquiries from the last 7 days related to delivery," and the AI automatically configures property filters and keyword filters.
• Reflected in the existing filter UI: The converted filters are reflected by populating values in the existing filter UI. You can also adjust them manually if needed.
• Automatic search execution: Once the filters are applied, the search runs immediately and the results are displayed.

Now, without having to think about which operator to use on which property, you can enter the conditions you want as-is and find your data right away. You can access data just like typing a search query.

A single phrase like "I need a metric like this" or "I want to find data like this" is enough. One line of natural language takes care of the configuration process, so you can focus on the decisions that matter.