Using the AI SDK to build Sitecore Stream's AI-powered brand aware assistant
By Steven Van ·
Marketers can brainstorm and request brand-aligned content, with responses that include text, images, and previews.
Marketers can ask Sitecore Stream for content ideas and visual asset suggestions aligned with their brand’s tone of voice. A case study from Vercel explains how Sitecore built the assistant using the Vercel AI SDK, connecting a Python backend with a Next.js frontend.
Brand conversations with visual responses
The assistant supports questions, brainstorming, and real-time content generation, responding with text, images, and content previews. Agent-based models are tailored to retain customers’ brand knowledge and tone, while the interface is designed for marketers regardless of technical expertise.
Streaming from backend to chat
Sitecore uses the AI SDK to manage data streaming between its Azure-hosted backends and Next.js frontend. The SDK’s useChat hook streams messages from the AI provider into the chat interface and manages input, messages, loading, and error states.
Alongside object generation, server-side rendering, and streaming, these components helped the team build interactive brand experiences without implementing all the underlying chat behavior themselves.