Your users have questions: Answer them directly inside of your - Bolt.new Workshop
Your users do not want to dig through docs, FAQs, or support pages just to get a clear answer. They want to ask a question in plain language and get a helpful, grounded response inside the app they are already using. In this hands-on Bolt.new Premium Customer Workshop, we'll build a conversational AI agent you can add directly into your app. We'll focus on one of the most useful real-world AI patterns: a support-style assistant that answers questions from your own documentation instead of guessing from generic model knowledge. *Presentation URL:* https://workshop-8-20-2026.bolt.host *Timestamps:* 00:00 Introduction & Welcome 04:58 Workshop Overview 11:58 Setting Up Mintlify 34:00 Building the AI Widget 50:39 Prompt Caching 54:55 Managing Knowledge Base 01:08:38 Wrap Up & Q&A *What we're building:* A conversational AI customer-support agent in Bolt. We'll create an in-app chat experience, connect it to a small knowledge base, retrieve the most relevant source material for each question, and generate grounded answers that users can trust. We'll focus on the practical pieces that make an AI agent feel useful in a real product: ingesting docs, structuring private knowledge, retrieving relevant context, writing safer prompts, showing source-aware answers, handling fallback responses, and embedding the assistant into an existing app flow. The goal is not just to add a chatbot, but to build a shippable support experience that helps users get unstuck faster. *Outcomes:* By the end of this workshop, you'll understand: How to design a conversational AI agent around a clear user problem and support workflow How to structure documentation or FAQ content so an agent can retrieve useful context How retrieval-augmented generation helps keep answers grounded in your own data How to build an in-app chat interface users can ask questions from directly How to think through common support-agent edge cases like missing answers, unclear questions, and sensitive information *Skills:* AI agents, conversational interfaces, retrieval-augmented generation, knowledge bases, support workflows, embeddings, prompt design, databases, source-grounded answers, app integration. Difficulty: Intermediate These workshops are designed to be stand-alone, so there is no need to attend prior sessions to get value from this one.