# Sprinklr MCP

Sprinklr MCP (Model Context Protocol) enables AI assistants to retrieve and analyze governed Sprinklr data through natural-language requests. Instead of navigating dashboards or building reports, users can ask business questions and receive data-backed answers in a conversational format.

## How It Works

1. **Ask:** Use natural language to ask a business question. Examples:
  - What is driving negative sentiment this week?
  - Which campaigns missed their ROAS target?
  - Which social posts generated the highest engagement?
2. **Understand:** Sprinklr MCP routes the request to the appropriate analytics copilot and analyzes data based on the user's existing permissions and workspace access.
3. **Act:** Receive a structured, narrative response that can help you:
  - Identify trends and performance changes
  - Understand key drivers and business impact
  - Share findings with stakeholders
  - Support data-driven decisions


## What Sprinklr MCP Does

Model Context Protocol (MCP) is an open standard that enables an AI application to use capabilities from an external system through natural-language requests. Sprinklr MCP applies this approach to Sprinklr analytics, so teams can explore business questions from an MCP-compatible assistant without first building or navigating a dashboard.

The result is a new way to access the same governed intelligence: users ask a question, the assistant routes it to the relevant Sprinklr copilot, and the response returns as an explainable narrative that can be refined in conversation.

## Core Capabilities

**Ask questions in plain language**
Explore performance and trends without learning dashboard structures, query syntax, or tool names.

**Use one conversational entry point**
Reach listening, organic social, paid media, and customer service analytics through a single MCP connection.

**Receive decision-ready narratives**
Turn metrics into written summaries that explain what changed, why it matters, and where teams may need to focus.

**Refine results through conversation**
Narrow a time period, compare segments, change the level of detail, or request a different output without starting over.

**Combine Sprinklr context with other work**
Where the AI client supports it, bring Sprinklr intelligence together with documents, enterprise data, and other connected sources.

**Work within existing governance**
Sprinklr MCP is read-only. Responses follow the signed-in user's workspace and role permissions, and the connection does not create, update, or delete Sprinklr data.

## Four Copilots, One Experience

Each copilot focuses on a distinct analytics domain. This specialization helps the assistant route questions to the right capability and return a relevant answer.

| Copilot | What You Can Explore | Example Question |
|  --- | --- | --- |
| Insights Copilot | Public conversation, share of voice, sentiment, themes, and competitors. | Where is sentiment shifting this week, and what is driving it? |
| Social Reporting Copilot | Organic account and post performance, engagement, and reach. | Which posts drove the most engagement last month? |
| Ads Reporting Copilot | Spend, pacing, CPM, CPC, CTR, and ROAS. | Which campaigns missed their ROAS target? |
| Service Reporting Copilot | Case volumes, queues, handle times, and response times. | Where are service queues backing up today? |


## What Your Teams Can Achieve

**Brand and Communications**
Track narrative shifts, understand sentiment drivers, compare share of voice, and summarize emerging themes for stakeholders.

**Social Media**
Identify high-performing posts, compare channels or accounts, explain engagement patterns, and use recent performance to inform content planning.

**Paid Media**
Monitor budget pacing, compare efficiency metrics, identify underperforming campaigns, and highlight where optimization may be needed.

**Customer Care**
Review case and queue trends, compare handling and response times, and identify operational pressure points that need attention.

**Leaders and Cross-Functional Teams**
Request concise briefings across domains, connect performance signals, and share a common narrative without waiting for separate reporting workflows.

## From Question to Business Outcome

| Start With | Explore | Produce |
|  --- | --- | --- |
| A performance question | Trends, comparisons, drivers, and exceptions | An executive summary with key takeaways |
| A change in sentiment | Themes, sources, competitors, and time periods | A narrative brief with recommended areas to investigate |
| A campaign concern | Spend, pacing, efficiency, and return metrics | A prioritized list of campaigns that need attention |
| A service bottleneck | Volumes, queues, handle times, and response times | An operational summary for the service team |


## Example Prompts

- "Summarize brand conversation from the last 30 days. Highlight the three most important changes and suggest actions for the communications team."
- "Compare organic performance across our main social channels. Explain which content themes drove the strongest engagement."
- "Identify campaigns that are pacing over budget or missing their ROAS target. Group the findings by priority."
- "Compare this week's service volumes and handle times with last week. Highlight the queues that need attention."
- "Create a leadership-ready summary with key findings, business impact, and three follow-up questions."


## How to Get Better Answers

You do not need a special query language. A well-framed question simply gives the assistant enough business context to return a useful answer.

1. **State the business objective:** Explain whether you want to monitor risk, improve performance, brief leaders, or find opportunities.
2. **Define the scope:** Include the relevant brand, account, campaign, queue, market, channel, or team.
3. **Add a time frame:** Specify a period such as today, this week, last month, or the last 30 days.
4. **Request a comparison:** Ask for period-over-period, channel, campaign, competitor, or segment comparisons when useful.
5. **Describe the output:** Request a summary, table, ranked list, trend explanation, or stakeholder-ready brief.
6. **Continue the conversation:** Ask follow-up questions to add detail, change the format, or focus on a specific finding.


## Responsible and Governed by Design

- **Read-only access:** Sprinklr MCP reads and returns information. It does not provide a path to create, update, or delete content or configuration in Sprinklr.
- **Permission-aware answers:** Users receive information that their Sprinklr workspace and role already allow them to access.
- **Governed source data:** Answers draw on the same standard Sprinklr dashboards and analytics available to the user.
- **Human judgment remains essential:** Use AI-generated summaries as decision support. Review important conclusions, recommendations, and business actions in context.


Sprinklr MCP gives teams a conversational path to Sprinklr intelligence. By bringing listening, social, advertising, and service analytics into an MCP-compatible AI experience, it helps users move from a question to a governed, shareable narrative with less reporting friction.