Using contextual chat in Magnity
Contextual chat means the AI can “see” the materials you are working with, not just your prompt. Instead of chatting in a blank window, you can attach relevant content such as an email draft, a flow overview, an industry description tile, or other source documents so the response is grounded in your context.
This makes it easier to refine and adapt content while you build emails, landing pages, social posts, and content agent outputs. It is based on the steps shown in the feature demo and focuses on three main topics:
- Opening chat in any application and working with the content you are building
- Adding context by dragging in tiles and documents from the content agent
- Uploading or crawling additional sources to guide the conversation
By the end, you should understand how to use contextual chat to align content to a flow or an industry, ask more precise questions, and iterate without losing track of what the AI is responding to.
1. Opening chat and working in context
Contextual chat is available across applications, including emails, landing pages, social posts, and the content agent. When you create an email from a page in your content library, you typically start with a ready structure, such as headline, body text, CTA, subject line, and preheader.
When you want to adjust the output beyond the base version, open the chat panel and use it to ask questions about what you are building. Contextual chat lets you connect your questions to the actual assets and documents in your workspace.
2. Adding context with tiles and documents
In the content agent, you can build contextual tiles by pasting in relevant text and saving it for reuse. You can then drag those tiles directly into chat to give the AI a specific perspective to work from.
For example, you can drag an overview of a flow into chat and ask whether an email draft fits that flow, or you can state that the draft is email one in the flow and continue building from there. The AI will read the attached context before responding and may recommend tweaks based on where the audience is in the engagement journey.
You can also drag in an industry description tile and ask the AI to align the email to that industry. This is useful when you want an industry specific angle without tying the content to a particular persona.
Once the context is in place, you can keep refining the conversation, for example by asking the AI to focus on one area such as measurement that matters rather than rewriting the whole email.
3. Extending context with your own sources
If you need additional background, you can add your own sources to the workspace and use them in contextual chat. The demo highlights several options, including:
- Uploading your own documents, such as a campaign brief
- Crawling specific URLs
- Adding raw sources
After adding sources, drag them into the chat the same way you would use tiles. This allows you to keep the conversation grounded in the material you want the AI to reference, and makes it easier to create and test variants that are aligned with your audience and the context you provide.
