> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bhealthpractice.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Search the supplement catalog using Ask Sina AI

> Use Ask Sina to describe a client's health situation in plain language and receive an AI-curated product stack with per-product clinical rationale.

Ask Sina is AnyScript's AI-powered search that translates a plain-language description of your client's situation into a curated supplement recommendation. Instead of manually browsing therapy areas and outcomes, you describe what the client is experiencing and Sina maps that description to matching products — surfacing a ranked stack with a clinical explanation for each recommendation.

## Opening Ask Sina

In the catalog toolbar, click the **Ask Sina** button (shown with a sparkle icon, styled in violet). The standard search field is replaced with the Sina input bar.

## Describing your client's needs

Type a natural language description of what your client is dealing with. You can be as specific as you like — the more clinical detail you include, the more targeted the results.

<Steps>
  <Step title="Type your query">
    Describe the client's situation in the Sina input bar. For example:

    > *"perimenopausal client with insomnia and joint pain"*

    > *"40-year-old male, low energy, poor focus, high stress job"*

    > *"child with recurring respiratory infections and poor gut health"*
  </Step>

  <Step title="Submit the query">
    Press **Enter** or click the **Ask** button. Sina begins processing your query — this typically takes 5–10 seconds.
  </Step>

  <Step title="Review Sina's interpretation">
    An interpretation panel appears showing how Sina understood your query: the matched therapy areas, clinical outcomes, and extracted keywords. This lets you verify Sina is on the right track before reviewing products.
  </Step>

  <Step title="Review the Recommended Stack">
    Sina ranks the top products from the catalog and presents them as a **Recommended Stack** — up to 5 products — with a written rationale explaining why this combination addresses your client's needs.

    Hover over the **Why** badge on any card in the stack to read the per-product reason Sina selected it.
  </Step>

  <Step title="Review supporting products">
    Below the Recommended Stack, additional products matching the same therapy areas and outcomes appear under **Other supporting products**. These are broader matches that may complement the primary stack.
  </Step>
</Steps>

## Saving the recommended stack

After Sina returns a stack, two action buttons appear beneath the stack rationale:

<CardGroup cols={2}>
  <Card title="Save as bundle" icon="package">
    Creates a named product bundle (up to 6 products) that appears on your public practitioner profile. Useful for showcasing curated collections to clients.
  </Card>

  <Card title="Save as protocol template" icon="clipboard-list">
    Saves the full stack as a reusable protocol template that you can apply to client scripts. The template carries over all products and their serving details.
  </Card>
</CardGroup>

Both options open a dialog where you can name the bundle or template before saving.

## Clearing an AI search

To return to the standard catalog view, click **Clear** in the interpretation panel, or click **Exit AI** in the Sina toolbar. All AI-applied filters (therapy areas, outcomes) are removed and the catalog resets.

<Tip>
  Describe symptoms and lifestyle context rather than just a condition name. "perimenopausal client with insomnia and joint pain" returns more targeted results than "menopause" because Sina can match specific clinical outcomes alongside relevant therapy areas.
</Tip>

<Note>
  Sina applies therapy area and outcome filters as a union — a product qualifies if it matches either the therapy area or the outcome signals from your query. This prevents over-filtering in cases where your query maps to multiple clinical signals.
</Note>

<Warning>
  Ask Sina requires a query of at least 3 characters. Very short or vague queries will prompt you to add more detail.
</Warning>