Gainsight PX MCP Server Integration Overview [Beta]
This article explains the PX MCP Server, including its supported capabilities, use cases, limitations, and available MCP tools for accessing PX data through AI assistants such as Claude and ChatGPT.
Overview
The Gainsight Product Experience (PX) MCP Server is an integration that connects your PX subscription with large language models (LLMs) such as Claude and ChatGPT. The PX MCP Server enables you to interact with PX using natural-language prompts to access product experience insights, user and account information, feature adoption metrics, engagement analytics, survey feedback, and other contextual PX data from supported AI assistants.
With the PX MCP Server, you can retrieve:
- Product usage and event activity data, including page views, sessions, custom events, and engagement interactions
- User and account information, including attributes, feature usage, preferences, and behavioral history
- Feature and module adoption metrics
- Engagement performance data including views, completions, errors, collisions
- Segment information
- In-App Hub (KC Bot) and article information
- Survey responses and NPS feedback
- Subscription and administrative metadata
This document provides an overview of the PX MCP Server, including its capabilities, supported use cases, limitations, and available tools. For more information on setup and configuration instructions, refer to the Connect Gainsight PX to LLMs Using MCP article.
How Teams Use PX MCP
The PX MCP server supports teams across Customer Success, Product, Growth, Operations, Onboarding, and Support by enabling conversational access to PX product experience data and insights within AI-assisted workflows.
| Team | Common Use Cases |
|---|---|
| Customer Success | Prepare for QBRs, track product adoption by account, and identify product-led health and expansion opportunities |
| Product Management | Understand adoption trends, identify friction points, and validate feature launches |
| Growth | Identify under-engaged segments and engagements with low view counts |
| Onboarding and Implementation | Track adoption milestones and onboarding engagement view counts |
| Product Operations | Audit instrumentation, validate event flows, and review segments and engagements |
| Support | Review user activity and help article coverage |
| UX Researchers | Combine usage data with survey feedback to identify insights |
| Data and Analytics | Export event data to CSV, audit attribute schemas, and build ad hoc product metrics |
| AI Agent Builders | Embed PX intelligence into automated workflows |
Supported PX MCP Capabilities
The current PX MCP release supports read-only access to PX data. You can retrieve and analyze PX information through supported AI assistants, but cannot create, update, or delete PX entities or modify PX data.
| Capability | Availability |
|---|---|
| Users | Supported |
| Accounts | Supported |
| Engagements | Supported |
| Features | Supported |
| Events | Supported |
| Surveys | Supported |
| In-App Hub (KC Bot) | Supported |
| Articles | Supported |
| Admin | Supported |
| User Preferences | Supported |
| Segment | Supported |
| External Segments | Not Supported |
Recommended Models
For reliable PX MCP responses, Gainsight recommends using the following models and effort levels:
| AI Application | Recommended Model | Minimum Recommended Model |
|---|---|---|
| Claude | Opus 5 | High Effort | Sonnet 5 | Medium Effort |
| ChatGPT | GPT-5.6 Sol | Medium Effort | GPT-5.6 Terra | Medium Effort |
Gainsight recommends to avoid using the Claude Haiku model with PX MCP, as it may produce unreliable responses.
Example Use Cases and Prompts
Use the PX MCP Server to retrieve and analyze PX data using natural-language prompts in supported AI assistants. The following examples demonstrate common product analytics, engagement, segmentation, feedback, and user intelligence workflows.
You can refine prompts with additional filters, grouping, and other criteria based on your analysis requirements. PX MCP also supports complex questions that combine multiple conditions and analytics dimensions.
Product Usage
| Use Case | Example Prompt |
|---|---|
| Identify top features based on account usage of a particular feature or module | Show me the top 10 features used by accounts that have used [feature or module name] more than 30 times in the last 7 days. |
| Visualize top feature usage over time | Create a dashboard visualization of the top 10 features in the last 7 days, with each feature displayed as a timeline widget for that period. |
| Understand active usage trend | Show me MAUs of last 6 months and provide insights accordingly |
| Pull feature adoption stats for any feature over a date window | What was the adoption of [Feature Name] last week vs the week prior? |
| Identify users who have or have not used a specific feature | List users from [Account Name] who have NOT used [Newly launched Feature Name] in the last 30 days. |
| Compare feature adoption before and after a launch | Show adoption of [Feature Name] for the 30 days before [Launch Date] vs the 30 days after. |
| Detect drop-offs, low usage patterns, or sudden spikes in activity | List users at [Account Name] whose last seen date is more than 30 days ago. |
| Export list of features as CSV | Get me all the list of features in the product [Product Name] used in last 7 days |
Account and User Intelligence
| Use Case | Example Prompt |
|---|---|
| Identify top accounts by Core Feature Usage | Show me the top 25 accounts by Core Feature Usage and export the results as a CSV file. |
| Identify accounts based on Product Score | Show me the accounts with a Product Score greater than 50, along with their Product Score, Active Users, Active User Coverage, and Stickiness. |
| Identify users with rage clicks by account | Show me the users, grouped by account, who had rage clicks in the last 30 days. |
| Understand feature or module adoption by accounts | Show me the feature adoption of [Feature Name], [optional Feature 2 Name], [optional feature 3 Name] by account. |
| Understand feature or module adoption by users | Show me the feature adoption of [Feature Name], [optional Feature 2 Name], [optional feature 3 Name] by user. |
| Identify at-risk customers | Show me the accounts at risk from the last 6 months. |
| Identify active users for an account | List the top 10 users from [Account Name] who have used [Newly launched Feature Name] in the last 30 days. |
| Identify inactive users for an account | List the bottom 20 users from [Account Name] who have NOT used [Newly launched Feature Name] in the last 30 days. |
| Find the top features used by an account | List the top 10 features used by [Account Name]. |
| Compare feature usage across accounts | List accounts in descending order of usage of [Feature Name] in the last 7 days. |
| Export lists of users | Fetch all users Note: The CSV export file generated by PX MCP is available via URL for 15 minutes. |
| Export lists of accounts | Export all accounts Note: The CSV export file generated by PX MCP is available via URL for 15 minutes. |
Engagements
| Use Case | Example Prompt |
|---|---|
| Engagement Performance Analysis | Provide a complete interaction breakdown of [engagement name] in the last 30 days , how many views and CTA clicks did it receive? |
| Find users who saw an engagement but did not adopt the target feature | Find users at [Account Name] who saw onboarding engagement [Engagement Name] but never used [Feature Name]. |
| List active engagements of a specific type with appropriate sorting | List all active engagements for content type GUIDE, sorted by most engagement view counts. |
Segmentation and Audience Targeting
| Use Case | Example Prompt |
|---|---|
| Identify which users or accounts currently match a given segment | Which users/accounts matched segment [Segment Name] in the last 30 days? |
| Cross-reference segment members with feature usage or event activity | Find users in the [At-Risk Segment] who also used [Pro Feature] in the last 30 days. |
Feedback, Surveys, and Research
| Use Case | Example Prompt |
|---|---|
| Identify detractors to target an in-app engagement or webinar | Find me the users who responded to any NPS survey as a detractor (score 0–6) in the last 60 days. Who qualifies? |
| Identify accounts who rated neutral to strongly agree or excellent | Show me the list of accounts who have responded three or more stars to any rating survey in last 30 days in product [product name] |
| Query survey responses over a date window | Summarize the survey responses from the last 30 days. |
| Flag accounts with concentrated negative sentiment | Which accounts have the highest count of negative survey responses in the last 60 days? |
| Triangulate survey sentiment with the responder's product usage | Cross-reference negative survey responses in the last 1 week with each responder's session and feature match activity over the last 1 month |
In-App Hub Analytics
| Use Case | Example Prompt |
|---|---|
| In-App Hub performance | For In-App Hub [In-app hub name] show Opens, Clicks, Users, Accounts, Interactions, Engagement Views, Article Views, Web Link Views, Searches , Task List views from April 1st to August 30, 2026 |
| Identify product areas with high traffic but no In-app Hub | List the top 20 URLs by pageView count in the last 30 days, and flag which ones are not in scope for any active In-App Hub. |
| List In-App Hubs configured for a product | List all In-App Hubs configured for product [Product Name]. |
Events
| Use Case | Example Prompt |
|---|---|
| Advanced Custom Event and property querying | How many [custom event name] events with [property name] had a value of [number] or more? Were those by new or returning customers? Show count per new Customer value in the last 30 days. |
| Query standard event activity | List the most recent 100 formSubmit events across all users. |
| Analyze custom event activity by property | Group custom events of name [Event Name] by [property field] for the last 30 days. |
| Validate event flow for a specific user | List all custom events fired by user [identifyId] in the last 24 hours. |
PX MCP Limitations
Review the following limitations before deploying or using the PX MCP Server:
| Limitation | Description |
|---|---|
| Read-only access | The current PX MCP release supports read-only access to PX data. You can retrieve and analyze PX information, but cannot create, update, or delete PX entities. |
| Permission-scoped access | The data and capabilities available through the PX MCP Server depend on the permissions associated with your user permissions. |
| Historical data range | Historical event analysis may be subject to PX event data date-range limitations. Large historical queries may require smaller time-range requests. |
| PX UI dependency for configuration tasks | Tasks that require visual configuration, governance, or manual review, such as engagement authoring or segment editing, must still be completed in the PX application. |
| Engagement lifetime views | Lifetime views are supported for a single engagement. When retrieving lifetime views across all engagements, data is limited to the past one year. |
Available MCP Tools
The PX MCP Server includes a set of underlying tools that enable supported AI assistants to retrieve and analyze PX data. Most users can interact with PX using natural-language prompts without directly referencing these tools. The following reference is provided for advanced users who want additional visibility into the available MCP capabilities and supported operations.
PX MCP v1.1 Tools
The following 16 tools are available with PX MCP v1.1 and include an optimized toolset for retrieving and analyzing PX data.
| Tool | Parameters | Purpose |
|---|---|---|
| px_text_to_query [New] |
question (required), propertyGroupId, propertyId, executeQuery, exportAsCsv, conversationContext, clarification | Query PX product analytics using natural language, across feature and module adoption and trends, engagements (guide, dialog, slider, email) Surveys and In-app Hubs performance users and accounts, health metrics etc. |
| px_list_products | - | List products in the subscription along with their product keys. |
| px_list_features | searchName, filter, propertyKey, pageNumber, pageSize, fetchAll, exportAsCsv | List features tracked in a product, or search for a feature by name. |
| px_list_engagements | searchName, propertyKey, contentTypes, pageNumber, pageSize, fetchAll, exportAsCsv | List engagements with their name, status, and type, or search for an engagement by name. |
| px_list_segments | productKey, searchTerms, pageNumber, pageSize, fetchAll, exportAsCsv | List segments defined in a product. |
| px_list_kc_bots | productId, pageNumber, pageSize, fetchAll, exportAsCsv | List In-App Hubs configured in a product. |
| px_get_model_attributes | type (required) | List available user or account fields, including custom attributes. |
| px_get_subscription | - | Fetch subscription details, including all products and their product keys. |
| px_get_account | accountId (required) | Fetch a single account record by ID. |
| px_get_user | identifyId (required) | Fetch a single user by identifyId. |
| px_get_feature | featureId (required) | Fetch a single feature by ID. |
| px_get_engagement | engagementId (required) | Fetch a single engagement by ID or name. |
| px_get_segment | segmentId (required) | Fetch a single segment by ID or name. |
| px_get_kc_bot | kcId (required) | Fetch a single In-App Hub by ID or name |
| px_get_engagement_scheduler [New] |
engagementId (required) | Fetch when an engagement was scheduled to run and the lifetime date range used to report on it. |
| px_get_custom_events | eventName, filter, dateRangeStart/End, sort, aggregation, fetchAll, exportAsCsv, scrollId, pageSize | Query custom events sent to PX over a selected date range, with optional grouping and totals. |
PX MCP v1.0 Tools
The following tools are specific to PX MCP v1.0 and are provided for reference for users who are using the older version.
| Tool | Parameters | Purpose |
|---|---|---|
| px_list_accounts | filter, sort, aggregation, fetchAll, csvExportPath, scrollId, pageSize | List accounts with filtering on industry, employee count, location, and custom attributes. |
| px_get_account | accountId (required) | Fetch a single account record by ID. |
| px_list_users | filter, sort, aggregation, accountName, accountFanoutOffset, fetchAll, csvExportPath, scrollId | List users with support for account-based filters and accountName shorthand. |
| px_get_user | identifyId (required) | Fetch a single user by identifyId. |
| px_list_features | propertyKey, filter, fetchAll, csvExportPath, pageNumber, pageSize | List product features with support for ID-based filtering. |
| px_get_feature | featureId (required) | Fetch a single feature by ID. |
| px_get_feature_adoption | featureId, propertyKey, dateRangeStart (required), dateRangeEnd | Feature adoption statistics over a selected date range. |
| px_get_events | eventType (required), filter, dateRangeStart/End, sort, aggregation, fetchAll, csvExportPath, scrollId | Query standard PX event types over a selected date range. |
| px_get_custom_events | eventName, filter, dateRangeStart/End, sort, aggregation, fetchAll, csvExportPath | Query custom events with attribute-level grouping. |
| px_list_engagements | contentTypes, filter, sort, fetchAll, csvExportPath, pageNumber, pageSize | List engagements filtered by content type and status. |
| px_get_engagement | engagementId (required) | Fetch detailed configuration information for a specific engagement. |
| px_list_segments | fetchAll, csvExportPath, pageNumber, pageSize | List available PX segments. |
| px_get_segment | segmentId (required) | Inspect a segment definition. |
| px_list_kc_bots | productId, fetchAll, csvExportPath, pageNumber, pageSize | List In-App Hubs (Knowledge Center bots), optionally filtered by product. |
| px_get_kc_bot | kcId (required) | Fetch configuration information for a specific In-App Hub (KC Bot). |
| px_get_survey_responses | filter, dateRangeStart/End, sort, aggregation, fetchAll, csvExportPath, scrollId | Query survey responses with date and attribute filters. |
| px_get_model_attributes | type='user' or 'account' (required) | Inspect standard and custom attribute schema information for users or accounts. |
| px_list_products | (none) | List products in the subscription along with their product keys. |
| px_get_subscription | (none) | Fetch subscription metadata for the authenticated PX account. |
The following tools currently only support filtering for the first name, last name, email, identifyID, accountID, and account Name attributes:
| Tool | Parameters | Purpose |
|---|---|---|
| px_get_feature_accounts_export | analyticsDataRequest (required), searchTerms, csvExportPath, useBigQuery | Exports account-level feature adoption data as a CSV export. |
| px_get_feature_audience_export | analyticsDataRequest (required), searchTerms, csvExportPath, useBigQuery | Exports user-level feature adoption data as a CSV export. |
| px_get_active_users | analyticsDataRequest (required), useBigQuery | Returns active-user trends over time for the selected property and date range. |
| px_get_accounts_at_risk | analyticsDataRequest (required), sortColumnName, sortOrder, useBigQuery | List of accounts flagged as at-risk based on product health score. |
| px_get_top_bottom_features | analyticsDataRequest (required), accountId (required), type, direction, limit, rankedBy, useBigQuery | Top or bottom performing features/modules by usage or unique users. |
| px_get_top_bottom_users | analyticsDataRequest (required), accountId, direction, limit, useBigQuery | Most or least active users ranked by visit count across accounts or within a specific account. |
| px_get_feature_account_table | analyticsDataRequest (required), searchTerms, useBigQuery | Provides account-level feature adoption metrics sorted by feature engagement. |
| px_get_feature_audience_table | analyticsDataRequest (required), searchTerms, pageSize, pageNumber, sortColumnName, sortOrder, useBigQuery | Returns paginated user-level feature engagement data for drill-down analysis. |
| px_get_feature_widget_timeseries | analyticsDataRequest (required), size, forceUpdate, useBigQuery | Returns feature usage trends over time grouped by feature. |
| Net Promoter®, NPS®, NPS Prism®, and the NPS-related emoticons are registered trademarks of Bain & Company, Inc., NICE Systems, Inc., and Fred Reichheld. Net Promoter ScoreSM and Net Promoter SystemSM are service marks of Bain & Company, Inc., NICE Systems, Inc., and Fred Reichheld. |