# Serasa Experian Credit Monitoring AI Agent Connect

> Serasa Experian Credit Monitoring MCP lets your AI agent manage the full credit monitoring lifecycle. Use it to watch specific CPFs or CNPJs for negative annotations, score fluctuations, and legal filings. You can build custom rules, manage large batches of documents, and set up webhooks so your agent receives real-time alerts instead of constantly polling for updates.

## Overview
- **Category:** finance
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_bTJXhr6tqdha0q5ju4OthHul4auhOTdCzfecFXZx/ai-agent-connect
- **Tags:** serasa, experian, credit-monitoring, monitora, alerts, webhooks, brazil, cpf, cnpj, risk-assessment, collections

## Description

You can now give your AI agent direct access to Serasa Experian's monitoring API. Instead of manually checking for credit events, you can instruct your agent to watch specific documents for changes in credit scores, bankruptcy filings, or notary records. The MCP handles the heavy lifting of authentication and token management, so you can focus on the logic of your risk workflows. You can create rules that target specific variables, such as the HAT3 or H5RC score families, and set custom intervals for when those scores should trigger an alert. For large-scale operations, you can add or remove up to 1,000 documents in a single batch. When a watched attribute changes, your agent can either pull the latest alerts or wait for Serasa to push them directly to your system via webhooks. This turns your AI from a simple chatbot into an active participant in your credit risk and collections operations.

## Tools

### add_monitored_documents
Adds CPFs or CNPJs to an existing monitoring rule. It returns a response separating accepted documents from those that failed with specific error codes.

### create_monitoring_rule
Creates a new rule to watch for specific credit events. You can define the document type, variable IDs, and alert emails, or set custom intervals for score monitoring.

### delete_webhook
Removes a registered webhook by its ID to stop alert delivery to those specific endpoints.

### list_alerts
Retrieves credit event alerts. You can filter these by document, rule name, or status, and use it to triage your unread alerts.

### list_monitored_documents
Shows all documents currently under monitoring. It includes details on their rules, inclusion dates, and current status.

### list_monitoring_rules
Lists all monitoring rules created within your organization, including the ability to filter for inactive rules.

### list_monitoring_subjects
Provides a list of available monitoring subjects like negative annotations or score families. Use this to find the numeric variable IDs needed for rule creation.

### list_notifications
Lists operational system notifications, such as updates on rule activation or document processing status.

### list_webhooks
Shows all currently registered webhook endpoints used for receiving pushed alerts.

### mark_alerts_read
Marks specific alerts as read to clear them from your unread queue.

### register_webhook
Sets up new webhook endpoints so Serasa can push credit alerts directly to your system.

### remove_monitored_documents
Stops monitoring for specific CPFs or CNPJs. The documents will stop generating new alerts but their history remains available.

### summarize_alerts
Aggregates alert data by document. It provides total and unread counts for a specific CPF or CNPJ.

### update_monitoring_rule
Modifies an existing rule. You can rename it, change the watched variables, or pause it by setting it to inactive.

## Prompt Examples

**Prompt:** 
```
Set up monitoring for CNPJ 33.014.556/0001-96: watch for negative annotations and bankruptcy, sending alerts to risk@company.com.
```

**Response:** 
```
I created the rule "Negatives & Bankruptcy" (id 42, type CNPJ) with the variables for negative annotations and bankruptcy, and added the CNPJ to it. The document is now monitored; alerts will be sent to risk@company.com.
```

**Prompt:** 
```
What happened with the monitored portfolio this week? Just give me the highlights.
```

**Response:** 
```
I aggregated the alerts per document. 5 of the 120 monitored documents had events this week: 3 new negative annotations, 1 notary record, and 1 score drop below the configured interval. The most affected is CNPJ 12.345.678/0001-90 with 2 events, both unread.
```

**Prompt:** 
```
Register a webhook so alerts get pushed to our collector at https://collector.example.com/serasa.
```

**Response:** 
```
I registered the webhook (id 7): the AUTH endpoint at https://collector.example.com/serasa/auth validates delivery, and ATTRIBUTE_NEGATIVE points to https://collector.example.com/serasa/negative. Serasa will now push negative-annotation alerts there instead of only being pollable.
```

## Capabilities

### Automated Rule Creation
Your agent can define specific credit event triggers for any CPF or CNPJ.

### Real-time Alert Delivery
The agent uses webhooks to receive instant notifications when credit attributes change.

### Bulk Document Management
You can add or remove up to 1,000 documents at once through your agent.

### Portfolio Triage
The agent can aggregate and summarize alerts to highlight the most critical credit events.

### Score Interval Monitoring
Your agent can watch for specific score ranges using custom interval configurations.

## Use Cases

### Real-time Credit Risk Alerts
An agent monitors a list of corporate clients and immediately notifies the risk team when a bankruptcy filing is detected.

### Automated Portfolio Auditing
A system uses the agent to summarize all credit events across a large group of CPFs every week.

### Dynamic Rule Management
A user instructs an agent to pause monitoring for a specific group of documents without deleting the underlying rules.

### Webhook Integration
A company connects their internal collector to Serasa via the agent to ingest credit alerts directly into their database.

## Benefits

- Replaces manual polling with push-based webhook notifications.
- Handles authentication and token refreshing automatically.
- Allows for batch processing of up to 1,000 documents per request.
- Enables granular monitoring of specific score families and variable IDs.

## How It Works

The MCP acts as a bridge between your AI client and the Serasa Experian API.

1. Connect your Serasa Experian Client ID and Secret to the MCP.
2. The MCP handles authentication and manages bearer tokens automatically.
3. Your agent calls tools to create rules or add documents to the monitoring list.
4. Serasa monitors the subjects and sends alerts via webhooks or makes them available for polling.
5. Your agent processes these alerts to trigger business actions or summaries.

## Frequently Asked Questions

**How do I add multiple documents at once?**
You can use the add_monitored_documents tool, which accepts a JSON array or a comma-separated list of CPFs or CNPJs.

**Can I monitor specific credit score ranges?**
Yes, when creating a rule, you can use the variables_config parameter to set custom interval ranges for score families.

**Does the MCP handle authentication?**
Yes, the MCP authenticates using your Client ID and Secret, then manages and refreshes the bearer token for you.

**How do I get notified immediately when a credit event occurs?**
You can use the register_webhook tool to set up endpoints that Serasa will use to push alerts directly to your system.

**What happens if I remove a document from monitoring?**
If you use the remove_monitored_documents tool, the document stops generating new alerts, but its historical alert data remains accessible.

**What is the difference between an alert and a notification?**
An **alert** is a credit event on a monitored document — a new negative annotation, a notary record, a bankruptcy filing, or a score crossing the configured interval. A **notification** is an operational event about the monitoring itself: rule created or activated, document batch processed, documents excluded, report downloaded. Alerts are the business signal; notifications are the audit trail.

**How do I know which variable ids to put in a rule?**
Call `list_monitoring_subjects` first — it returns each subject (Negative, SpcNegative, Bankruptcy, NotaryRecord, HAT3, …) with its numeric id, the document types it applies to, and, for score subjects, the allowed custom interval range. The `variables` you pass to `create_monitoring_rule` are exactly those ids, as a JSON array or a comma list like "1,2,3".

**Why did adding documents return a 207 with some rejections?**
A 207 means the batch was partially accepted: some documents went through, others were rejected. The result separates them into accepted and rejected lists with per-document error codes — the most common is a document already monitored by that rule. Only genuinely invalid input or an auth failure fails the whole call.