The Appcues MCP connector lets AI assistants pull your survey data directly from Appcues and turn it into visual dashboards — score distributions, sentiment breakdowns, theme analysis, and response timelines — without writing code or exporting CSVs. Any AI tool that supports MCP can connect to Appcues this way, including Claude, ChatGPT, Cursor, and others.

This guide walks through the workflow using Claude, but the core pattern — connect the MCP, ask for your data, request visualizations — applies to any MCP-compatible AI assistant.

This works with any Flow that collects survey responses, including NPS surveys, feedback forms, and multi-question surveys built with the Appcues form components.

Prerequisites

• An AI assistant that supports MCP connections (this guide uses Claude on a Pro, Team, or Enterprise plan)
• An Appcues account with at least one published Flow that collects survey responses
• The Appcues MCP connector enabled in your AI tool (see Use the Appcues MCP Server with your AI tools for the complete setup)

Pull survey results

Ask your AI assistant to retrieve results from a specific Flow. Be specific about the Flow name so it can find the right one.

Example prompts:

• "Summarize the survey results from my Flow called [Feature release] Onboarding Feedback Survey"
• "Pull the survey responses from my NPS survey and tell me what people are saying"
• "Get all survey responses from my post-launch feedback Flow between January and March 2026"

The assistant searches your Appcues account for the matching Flow, retrieves the survey responses, and summarizes the results — including scores, open-ended feedback, and response metadata.

Generate visual dashboards

Once your assistant has the survey data, ask it to visualize the results. Most MCP-compatible AI tools can generate charts directly in the conversation — the exact output format varies by provider.

Example prompts:

• "Build a dashboard to visualize these survey results"
• "Show me a chart of the score distribution"
• "Create a visual breakdown of the feedback themes"
• "Chart the response timeline by month"

What you can generate

• Score distributions — bar charts showing how many respondents gave each rating
• Sentiment breakdowns — pie or donut charts grouping responses into negative, neutral, and positive
• Feedback theme analysis — ranked bars showing the most common topics in open-ended responses, with severity indicators
• Response timelines — line or area charts showing when responses arrived over time
• Summary metric cards — at-a-glance stats like average score, total responses, and top issues

You can ask for all of these at once ("build a full dashboard") or request individual charts one at a time. 

Customize and iterate

The visualizations are conversational — ask follow-up questions to refine them.

Examples:

• "Filter out internal test responses and recalculate"
• "Break down the scores by company"
• "Highlight which respondents mentioned bugs"
• “Add the individual responses as an expandable list below the charts”

Tips for better results

• Name your Flows clearly. The assistant searches by name, so descriptive titles like "[Feature release] Onboarding Feedback (Q1 2026)" work better than "Survey v2."
• Include date ranges. If your Flow has been live for a long time, specify the period you care about: "Get responses from the last 30 days."
• Ask for theme analysis on open-ended questions. AI assistants can read the full text of each response and group recurring topics — this is where the MCP adds the most value over manual CSV analysis.
• Combine with other data. You can ask the assistant to cross-reference survey results with engagement analytics from the same Flow, like how many users saw the survey vs. how many responded.

Limitations

• Survey data is retrieved as a snapshot — it doesn't update in real time as new responses come in. Start a new conversation or ask the assistant to re-fetch for the latest data.
• Very large response sets (1,000+) may be truncated. Use date range filters to narrow the results.
• Visualizations render in the conversation and can often be copied as code (React, HTML) but aren't embeddable directly in other tools without additional setup. Exact export options depend on your AI provider.
• The Appcues MCP is read-only — your AI assistant can pull data but can't modify your Flows or survey settings.