> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wizflow.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Data Models in the Assistant
> Have the Wizflow assistant map a flow to a data model, build a flow from one, or create a custom data model
The [AI Assistant](/ai-assistant/overview.md) can now work with [data models](/features/data-model-editor.md). Ask it to connect a flow's questions to the household or financial plan data model, build a new flow from the fields of a data model, or make a custom data model for something like KYC. Nothing changes until you have looked at the proposal and approved it.
Building a flow from the household data model, with a calculation and illustrations
***
## What it can do
> **Map an existing flow**
>
> Connects the questions in a flow you already have to the matching fields of the household, financial plan or another standard data model.
> **Build a flow from a data model**
>
> Writes the questions for the fields you name, such as the user section of the household model, and maps each answer to its field.
> **Create a custom data model**
>
> Designs a data model for data no standard model covers, for example the answers in a KYC flow, and maps the flow to it.
> **Change data model settings**
>
> Turns settings on or off for the flow's data model, such as saving only when the flow is completed, or keeping the data when the chat is cleared.
The assistant can also save a [variable](/features/variables.md) to a field instead of a raw answer. Ask for an annual salary and have it stored as a monthly figure, and it adds the variable that does the conversion and maps that.
***
## Using it
**1. Open the assistant**
Open the assistant in the flow builder.
**2. Ask for the data model work**
For example: "Map the questions in this flow to the household and financial plan data models", or "Create a KYC flow and a custom data model to store the answers."
**3. Preview the proposal**
The assistant shows an approval card. Click **Preview** to see a table of the fields and the questions mapped to them.
**4. Approve**
Click **Approve** to make the changes. The data model and its mappings are then in the flow like any you set up by hand.
Creating a custom data model and changing its settings
> [!TIP]
>
> Combine it with the other skills in one prompt. "Create a flow from the household data model, one section per part, an illustration for each section, and run the household calculation at the end" builds the questions, the mappings, the [images](/ai-assistant/asset-generation.md) and the trigger together.
> [!NOTE]
>
> The assistant can't delete a data model or edit a custom data model once it exists. Do that in the [data model editor](/features/data-model-editor.md).
***
## Next steps
> **[Data Model Editor](/features/data-model-editor.md)**
>
> See and edit the data models the assistant made
> **[Financial Plan Model](/features/financial-plan-model.md)**
>
> The shared plan model the assistant can map to
> **[Themes in the Assistant](/ai-assistant/themes.md)**
>
> Have the assistant style the flow too
> **[Autofill](/features/autofill.md#profiles-from-the-assistant)**
>
> Have it make test profiles to click through the new flow
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.