# Quick start > The different entry points for starting a calculation: from the dataset panel, viz settings, or the result table, plus composing from an existing field and AI. You can start any of these from the **dataset panel**, **viz settings**, or the **result table**. This section focuses on creating metrics with **Calculation Builder** (GUI). For Formula mode and AQL, see the [AQL](/as-code/aql) docs. ## Create metrics from scratch Use this when you want to create a new metric with blank settings and set it up yourself. See [Available calculations](/docs/calculation-builder/available-calculations) for the full list of options. ## Create metrics from an existing one Use this when you want to build a new metric using an existing one as the input. **Example:** from `Total orders` metric, create a new metric `% of total orders`. ## Aggregate dimensions into metrics Use this when you want to summarize a dimension into a metric with a quick aggregation (e.g. count, sum, average, min, max...). **Example:** from the `Delivery attempts` dimension, apply 75th percentile to create a metric `P75 - Delivery attempts`. ## Duplicate existing metrics Use this when you want a variant of an existing metric with a small tweak. **Example:** duplicate `Total orders` into a new metric, then change its filter to `Country is Australia` to make `Total orders (Australia)`. ## Create metrics with AI Describe what you want in plain language and AI builds the metric for you, no point-and-click needed. You can easily verify it's accurate by seeing how AI built it in the chat, and if you'd like to change anything, just open the metric in the calculation builder. **Example:** ask for `Ratio of Australia orders to total`. AI builds it as `Total orders (Australia)` ÷ `Total orders`, reusing those metrics if they exist or creating them if not.