Talk To Your Data Application

Talk To Your Data (TTYD) lets authorized Cleverbridge users analyze client performance data on supported AI platforms using natural language.

TTYD is a Cleverbridge analytics assistant for self-serve, insight-oriented data analysis. It gives authorized users natural-language access to Cleverbridge analytics data that they are already permitted to access through Cleverbridge reporting and analytics tools.

The application is designed for exploratory, multi-step, or insight-oriented analytics rather than simple one-off reporting. It can help users move from "what happened?" to "why did it happen?" and "what should we do next?" by combining relevant metrics, calculating derived fields, comparing periods, identifying unusual changes, and producing structured outputs such as tables, chart-ready datasets, bridge reports, subscription health reports, and action recommendations.

Users must access TTYD through an approved Cleverbridge AI workspace, connector, or integration and complete the applicable Cleverbridge authorization flow before using it.

Application usage

Target users

Primary users include:

• Customer Success and CSM teams
• Account management teams
• Revenue, subscription, and ecommerce analytics users
• Finance, ecommerce, marketing, sales, IT/engineering, and client-success stakeholders who need client-level performance analysis

When the application is used

AI platform will use the applicatoin if the users requests to analyze Cleverbridge client, revenue, subscription, product, country, payment, or performance data that is available to them through Cleverbridge analytics systems. Typical triggers include requests such as:

• Analyze revenue, MRR, ARR, churn, renewals, new business, subscriptions, active subscribers, ARPU, product performance, country performance, or payment performance for a specific Client ID.
• Create a revenue, subscription, payment, or financial report for a client over a specified period.
• Compare current-year results to last year, month-over-month performance, or other period-over-period changes.
• Identify trends, anomalies, spikes, drops, drivers, or likely causes in Cleverbridge performance metrics.
• Produce tables, chart-ready datasets, bridge reports, subscription health reports, or concise business summaries from Cleverbridge analytics data.
• Recommend concrete actions based on client performance data.

AI platform should not use the application if at least one of the following conditions are true:

• The user is asking a general analytics, finance, ecommerce, or subscription-business question that does not require Cleverbridge client data.
• The user does not provide enough information to identify the relevant client, period, metric, or dimension, unless the app can ask a clarifying question first.
• The requested data is outside the authenticated user's Cleverbridge permissions.
• The user asks for personal data, confidential client data, or restricted information they are not authorized to access.
• The request is better served by a quick standard report in existing Cleverbridge reporting tools and does not require deeper analysis, reconciliation, narrative explanation, or custom aggregation.
• The user asks the app to modify production data, trigger client-facing communication, or perform operational actions beyond analytics.

Primary use cases and expected outcomes

As a result of a request, the application can return:

• Concise natural-language answers
• Tables with one row per period or dimension
• Period-over-period and year-over-year comparisons
• Currency-aware outputs
• Chart-ready datasets
• Driver analysis
• Anomaly detection summaries
• Reconciliation checks where applicable
• Recommended actions based on findings
• Follow-up questions when the user request is underspecified

These are some of the primary use cases for the application.

Standard metric retrieval and comparison

Users can request reports that reproduce or summarize values otherwise available in existing analytics tools.

Possible prompt Create a Net Revenue monthly view for 2025 for Client nnn. Compare it with 2024 and output a table with one row per month, including Month, Net Revenue 2025, Net Revenue 2024, and YoY % change. Also create chart-ready data for a line chart with current year as a solid line and last year as a dashed line.

Expected results:

• Monthly net revenue table
• Year-over-year comparison
• Chart-ready data
• Short explanation of notable changes

Extended reporting and derived analysis

Users can combine multiple measures into one structured output with consistent deltas and reconciliation checks.

Possible prompt Create a monthly MRR bridge report for Client nnn for the last 6 months for all available currencies. Compare values month to month and give a total for the full period.

Output columns: Month, Starting MRR, New MRR, Expansion MRR, Contraction MRR, Churned MRR, Ending MRR, Net MRR Change.

Show currency symbols, round consistently, show negative values with a leading minus sign, and include chart-ready data that illustrates the report.

Expected results:

• Monthly bridge table per relevant currency
• Total period summary
• Chart-ready data
• Explanation of what drove MRR changes

Subscription health and anomaly analysis

Users can ask for broader analysis across subscription metrics and dimensions.

Possible prompt For Client nnn, run a monthly subscription health report for the last 12 months. Include MRR, ARR, Churn MRR, Renewals MRR, New Business MRR, Active Subscribers, and ARPU. Break down by Product, top 10 by MRR, and Country, top 5. Identify months with unusual Churn MRR spikes, quantify impact on total MRR, explain likely drivers such as product or country mix shifts, and recommend 3 concrete actions to reduce churn or improve ARPU.

Expected results:

• Monthly subscription health report
• Product and country breakdowns
• Anomaly detection and quantified impact
• Likely driver analysis
• Recommended actions

Open-ended performance advisory

Users can ask less structured questions and receive an analysis-style answer from available metrics.

Possible prompt What should I know about Client nnn performance over the last 12 months?

Expected results:

• Identification of relevant metrics
• Summary of key trends and risk areas
• Suggested follow-up questions or actions
• Permission prompts before accessing new datasets or tools where required

Handling data

Talk To Your Data (TTYD) retrieves and analyzes Cleverbridge analytics data only when an authorized user asks a question that requires that data. The application supports read-only analysis: it can retrieve available metrics, aggregate results, compare periods, identify patterns, and produce summaries or chart-ready outputs, but it cannot modify production data.

Data access is governed by the authenticated user's existing Cleverbridge permissions, client scope, and configured data access. TTYD does not create a separate data-access layer or expand what a user can see. If a user is not authorized to access a client, metric, dataset, or integration in Cleverbridge, that data is not available through the application.

Data sources

The data sources used by the application include:

• Cleverbridge analytical dataset — client-level analytical data used to monitor customer behavior, track revenue, and uncover business trends through Cleverbridge analytics and reporting tools.
• Report Builder — Cleverbridge's self-service reporting tool for creating reports using available dimensions, metrics, and filters.
• Available metrics and dimensions — the documented set of metric and dimension combinations supported by Report Builder.
• SCM Analyze / Data & Analytics — Cleverbridge's analytics area for analyzing and reporting in SCM, including dashboards, Report Builder, Insights, filters, comparison views, custom reports, and bookmarks.
• Subscription analytics — subscription-related analytics and customizable reports for metrics such as churn, MRR, ARR, renewals, new business, active subscribers, ARPU, and related KPIs.
• Multi-item analytics — analytics for subscription behavior involving multiple products, bundled offers, tiered subscriptions, and item-combination patterns.
• Payment analytics — payment-performance analytics for monitoring payment performance, revenue trends, refunds, chargebacks, authorization behavior, and related payment KPIs.
• Analyze payment performance — payment analytics for submitted and paid orders, payment-method performance, authorization declines, and recovered revenue.
• Insights — AI-powered analytical outputs such as forecasting, anomaly detection, and subscription KPI benchmarking.
• Snowflake data sharing, APIs, and Integration Hub — client-specific data available through the corresponding Cleverbridge data-sharing or integration setup, where configured for the authenticated user and client.

Data access

• Users access the app through an approved Cleverbridge AI workspace, connector, or integration.
• First-time users may need to connect the Cleverbridge TTYD app or connector and complete the applicable Cleverbridge authorization flow.
• Authentication may route through Cleverbridge SSO depending on the user's settings and the AI platform used.
• After authorization, users can start a conversation with the Cleverbridge TTYD app or connector enabled.
• Data access is always limited to data the authenticated user is already allowed to access in Cleverbridge analytics tools.

User setup flow

Users access the app or connector through a supported AI platform where it is available to their organization or account.

ChatGPT setup flow

1. Open ChatGPT.
2. Go to Apps.
3. Search for Cleverbridge or Talk To Your Data.
4. Select the app.
5. For first-time setup, click Connect.
6. Continue to Cleverbridge.
7. Complete SSO or Cleverbridge login if prompted.
8. After connecting, click Start chat to open a conversation with the app enabled and ask the first analytics question.

Claude setup flow

1. Open Claude.
2. In a new or existing chat, click the + button in the message composer.
3. Select Add connectors.
4. In the connector directory, open Connectors.
5. Select the Cleverbridge connector.
6. Review the connector details and click Connect.
7. If required, Claude opens the Cleverbridge authorization flow in a new browser tab. Click Approve to grant access.
8. Return to Claude, start a conversation and ask the first analytics question.

note

• The app is available only to users and organizations approved by Cleverbridge.
• Users only need to complete the full authorization flow the first time they connect.
• When accessing the agent for the first time, the user may be asked to approve access.
• Data access is limited to the authenticated user's Cleverbridge permissions.

Did you find this doc useful?