feat: add page about ai chat

chat.mdx

@@ -0,0 +1,43 @@
+---
+title: "AI Chat"
+description: "Learn how the AI Chat feature uses your docs as context to answer end-user questions"
+---
+
+<Info>
+  AI Chat is available on the Pro plan and up.
+</Info>
+
+## Overview
+
+The AI Chat feature, embedded in the search bar, allows end-users to get help via a chat interface instead of using full-text search or navigating manually through your documentation.
+
+<Frame caption="The AI Chat interface within the search bar.">
+  ![Screenshot of the AI Chat interface within the Mintlify search bar, showing a question and answer.](/images/ai-chat.png)
+</Frame>
+
+**Benefits:**
+
+- **For End-Users:** Allows them to ask specific questions about your product/service and receive instant answers sourced directly from your documentation.
+- **For Documentation Maintainers:** Helps reduce the load on your support teams by addressing common questions and improves engagement with your documentation.
+
+## How it Uses Your Documentation
+
+AI Chat uses only the content published on your docs as its knowledge base. We designed it to use your documentation as the single source of truth and avoid hallucinating answers that are not supported by your content.
+
+User queries are matched against the contents of your documentation for both semantic & full-text matches, which are then provided as context to a LLM.
+
+You can choose the specific Large Language Model (LLM) used for generating answers from a list of supported models within your Mintlify dashboard. This allows you to select the model that best fits your needs for performance, cost, or response style.
+
+It also takes into account the end-user's selected version and language and filters the sources it uses as context based on these factors.
+
+## Optimizing Your Content for AI Chat
+
+To ensure the AI Chat Assistant provides the best possible answers to end-users, consider these content strategies:
+
+- **Follow Search Best Practices:** Optimizing for traditional search is the first step. Ensure your documentation uses clear headings, logical structure, focused sections, and natural keyword inclusion. Refer to our [search guide](search.mdx) for details.
+- **Write Explicitly & Answer Questions:** Phrase content clearly and directly answer anticipated end-user questions. Instead of just describing features, explain _how_ to use them or what settings do. This gives the LLM direct context to formulate answers.
+- **Consider FAQ Pages:** Adding FAQ pages can provide pre-packaged Q&A pairs that the LLM can easily use.
+- **Use Summaries:** For long or complex pages, a brief summary paragraph at the start or end helps the LLM answer high-level questions accurately.
+- **Define Key Terms Clearly:** Ensure you define product-specific jargon, acronyms, or technical terms inline or in a glossary.
+- **Maintain Consistency:** Use consistent terminology for the same concepts throughout your docs to avoid ambiguity.
+- **Structure Data Clearly:** Use tables or definition lists for structured data like parameters or error codes, making them easier for the AI to parse accurately.