Chat With Your Research Using Strella's AI Chat Feature

Use Strella’s built-in AI Chat to ask questions, surface participant insights, find cross-project patterns, and generate stakeholder reports.

AI Chat is one of Strella’s most powerful features. Think of it as a built-in ChatGPT that knows everything your participants said — you can ask it questions, uncover patterns, generate reports, and save the results as highlight reels, all without leaving your project.

Access AI Chat

In the top navigation bar of any project page, click AI Chat to open the chat interface. AI Chat has two modes: Find Insights and Create a Report.

AI Chat is available on every project page. You don’t need to navigate away from your current view — just click the button in the top nav bar.

---

The two AI Chat modes

Find Insights

Use this mode to ask open-ended questions about your research and let AI search across all participant sessions to pull out relevant answers.

Ask a question

1. Select 'Find Insights' mode

   Make sure the Find Insights tab is active in the AI Chat interface. This is the default mode when you open chat.

2. Type your question and send it

   Ask anything related to your research — for example, “What frustrations did participants mention about the onboarding experience?” or “Were there any moments of surprise or delight?”

3. Review the response and participant quotes

   AI Chat returns a synthesized response. When the answer includes a participant quote, you’ll see a circular icon with the participant’s initials next to the relevant text. Click the icon to expand the exact quote, then click the quote itself to jump directly to that moment in the participant’s session.

Save an insight

When AI Chat returns a response that includes participant quotes, a Save insight button appears at the bottom of that message. Click it to save the insight to your Saved tab, where Strella automatically generates an associated highlight reel.

See Creating Highlight Reels and Insights Board for more on what happens after you save.

Find insights across multiple projects

1. Open AI Chat and look for the project selector

   In the chat interface, locate the option to select which projects to include in the search.

2. Select multiple projects

   Choose two or more projects to combine their participant data in a single conversation. This is particularly useful when a study has been split across multiple Strella projects and you want to identify themes that span all of them.

3. Ask cross-study questions

   Pose questions as you normally would. AI Chat searches across all selected projects simultaneously and surfaces patterns consistent across participant groups.

Cross-project chat is especially valuable for longitudinal studies, multi-cohort research, or when comparing responses from different participant segments run as separate projects.

Manage your chat history

The Explorer sidebar lists every AI Chat you’ve started, so you can return to a previous conversation, keep a working set of chats handy, or narrow the list to a single project. Starred chats are pinned to the top of the list, and the sidebar can be dragged to a custom width — your preferred width is remembered the next time you open the Explorer.

1. Star a chat from the row actions menu

   Hover over a chat in the sidebar and open the row’s actions menu (the icon at the end of the row). Click Star to mark the chat as a favorite. Starred chats display a star icon inline and are pinned to the top of the chat list so they stay visible as your history grows. Click Star again to unstar.

2. Filter by project

   Click the Filter icon in the sidebar header. Choose the Project field and select one or more projects to limit the list to chats scoped to those projects. The project options are drawn from the projects referenced by your existing chats.

   When you open the Explorer from a project page, the sidebar arrives pre-filtered to that project via a ?project=<slug> URL parameter, and the filter persists as you click between chats. Starting a new chat from the sidebar carries the active project filter into the new chat, so the conversation is scoped to the project you were already exploring.

3. Filter by favorite

   In the filter popover, choose the Favorite field and set it to is favorite to show only starred chats, or is not favorite to hide them. Combine this with the project filter to focus on the starred chats inside a specific project.

Drag the right edge of the sidebar to resize it. Use the collapse control in the sidebar header to snap it closed when you want more room for the chat view, then expand it again to return to your previous width.

Create a Report

Use this mode to generate a polished, structured report from your participant data — ready to share with stakeholders.

Generate a report

1. Select 'Create a Report' mode

   In the AI Chat interface, click the Create a Report tab to switch modes.

2. Describe the report you want

   Tell AI Chat what kind of report you need. Be as specific as helpful — for example, “Create an executive summary of key usability findings with supporting quotes” or “Generate a report on how participants responded to the new pricing page.”

3. Review the generated report

   AI Chat searches across all sessions in your selected project, pulls in relevant insights, themes, and participant responses, and produces a tailored report. Read through it and identify anything you’d like to refine.

4. Continue chatting to refine the report

   Reply in the chat to adjust the report — ask for more detail on a specific section, a different tone, additional participant quotes, or a shorter summary. AI Chat updates the report iteratively.

5. Save and share the report

   Once you’re satisfied, save the report. For full details on formatting options and sharing reports with stakeholders, see the Navigating Reports page.

Start with a broad prompt to get an initial structure, then follow up with targeted requests to add depth to specific sections. Treating it like a conversation produces better results than trying to specify everything in a single message.