> ## Documentation Index
> Fetch the complete documentation index at: https://neuroai-cf1c4abf.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Builder with Super Agent

## Build Custom AI Agents in Minutes - No Coding Required

A comprehensive guide to creating powerful custom agents using NeuroAI's intuitive Agent Builder and Super Agent system.

***

## Table of Contents.

1. [What is Agent Builder?](#what-is-agent-builder)
2. [Super Agent Overview](#super-agent-overview)
3. [How to Build Your Custom Agent](#how-to-build-your-custom-agent)
4. [Step-by-Step Creation Process](#step-by-step-creation-process)
5. [Agent Builder Features](#agent-builder-features)
6. [Real-World Examples](#real-world-examples)
7. [Best Practices](#best-practices)
8. [Troubleshooting](#troubleshooting)
9. [FAQs](#faqs)

***

## What is Agent Builder?

### Definition

**Agent Builder** is NeuroAI's no-code platform that allows you to create custom AI agents by simply describing what you want them to do. No programming knowledge required.

### Key Concept

Instead of writing code or complex configurations, you describe your agent's purpose, responsibilities, and behaviors in plain English (or your preferred language). The Agent Builder then creates a fully functional AI agent that can:

* Execute tasks autonomously
* Integrate with 700+ applications
* Make decisions based on rules you define
* Communicate with your team
* Handle complex multi-step workflows

### Who is it for?

* **Business Users:** No technical background needed
* **Team Leaders:** Automate departmental workflows
* **Entrepreneurs:** Accelerate business processes
* **Managers:** Delegate repetitive tasks to AI

***

## Super Agent Overview

### What is Super Agent?

**Super Agent** is NeuroAI's most powerful AI orchestrator that:

* Understands natural language descriptions of complex tasks
* Breaks down complex requirements into actionable steps
* Coordinates multiple AI models and tools
* Manages 700+ app integrations
* Executes workflows autonomously

### Super Agent Capabilities

| Capability                      | Description                                       |
| :------------------------------ | :------------------------------------------------ |
| **Natural Language Processing** | Understands complex descriptions in plain English |
| **Multi-Model Coordination**    | Uses best AI model for each task type             |
| **Tool Management**             | Controls 700+ application integrations            |
| **Decision Making**             | Makes autonomous decisions based on rules         |
| **Learning**                    | Improves performance over time                    |
| **Error Handling**              | Recovers from failures gracefully                 |
| **Scheduling**                  | Runs on schedule or triggered by events           |

### Why Super Agent?

The "Super" in Super Agent means it's:

1. **All-in-One:** Does what would normally require 5+ separate tools
2. **Intelligent:** Understands context and nuance
3. **Autonomous:** Works 24/7 without human intervention
4. **Adaptable:** Learns and improves over time
5. **Scalable:** Handles simple to extremely complex tasks

***

## How to Build Your Custom Agent

### The Simplest Possible Process

1. **Describe** what you want the agent to do
2. **Select** the apps it needs
3. **Set** any rules or limitations
4. **Deploy** and start using

That's it! No coding, no complex configuration.

### Example: Building a Sales Agent

**What you say:**

```
"I want an agent that:
- Searches for leads in companies in the tech industry
- Researches each company on the web
- Writes personalized emails to decision makers
- Tracks responses in Salesforce
- Follows up after 3 days if no response"
```

**What happens:**

1. Agent Builder creates your Sales Agent
2. It integrates with: Google Search, Email, Salesforce, Calendar
3. It runs every day automatically
4. It handles the entire sales process

**Result:** You get a fully functional sales team working 24/7

***

## Step-by-Step Creation Process

### Step 1: Access Agent Builder

**Where to find it:**

1. Log into NeuroAI dashboard
2. Click **"Create New"** → **"Custom Agent"**
3. Or go to **"Agent Builder"** from main menu

### Step 2: Name Your Agent

**Good agent names:**

* Sales Outreach Agent
* Customer Support Bot
* Research Assistant
* Content Creator
* HR Interview Coordinator

**Bad agent names:**

* Agent1
* Bot
* AI (too generic)

### Step 3: Describe Your Agent's Purpose

**What to include:**

```
My agent should:
1. [Main responsibility]
2. [Secondary responsibilities]
3. [Expected outcome]

Context:
- Who will use it: [Team/Department]
- When it runs: [Schedule/Trigger]
- What apps it needs: [List]
```

**Example for Marketing Agent:**

```
My agent should:
1. Create 5 social media posts per week
2. Schedule them for optimal engagement times
3. Monitor comments and replies
4. Report weekly performance metrics

Context:
- Who will use it: Marketing team
- When it runs: Every day at 2 PM
- What apps it needs: 
   - Twitter
   - LinkedIn
   - Facebook
   - Google Analytics
   - Slack (for reporting)
```

### Step 4: Select Applications

**How it works:**

1. Agent Builder shows recommended apps based on description
2. You can add/remove apps as needed
3. Click "Connect" to authorize each app
4. Choose what permissions the agent gets

**Permission Levels:**

| Permission  | What Agent Can Do | Example                       |
| :---------- | :---------------- | :---------------------------- |
| **Read**    | View data only    | Read emails, check calendar   |
| **Write**   | Create new data   | Send emails, create tasks     |
| **Delete**  | Remove data       | Delete old records            |
| **Execute** | Trigger actions   | Run workflows, schedule posts |

⚠️ **Best Practice:** Only grant permissions the agent actually needs.

### Step 5: Define Agent Behavior

**You describe:**

```
Rules and Constraints:
- Always [do this]
- Never [do that]
- If [condition], then [action]
- Before [action], [check this]

Quality Standards:
- Accuracy: [High/Medium/Low]
- Response time: [Immediate/Within 1 hour/Within 24 hours]
- Tone: [Professional/Friendly/Formal]
- Language: [English/Russian/Other]
```

**Example for Support Agent:**

```
Rules:
- Always respond within 1 hour
- Never promise things we can't deliver
- If issue is urgent, escalate to manager immediately
- Before responding, check knowledge base

Quality Standards:
- Accuracy: Very high (customer satisfaction critical)
- Response time: Within 1 hour
- Tone: Professional and empathetic
- Language: English
```

### Step 6: Set Up Triggers & Schedule

**Triggers (When agent runs):**

* **Manual:** When user clicks "Run"
* **Scheduled:** Daily at 9 AM, Weekly on Monday, Monthly on 1st
* **Event-based:** When new email arrives, new lead in Salesforce, form submitted
* **API:** When external system calls NeuroAI

**Example configurations:**

```
Sales Agent:
- Run every weekday at 9 AM
- Also run when new lead enters Salesforce

Support Agent:
- Run immediately when new support ticket arrives
- Run every 6 hours to check for unresponded tickets

Marketing Agent:
- Run daily at 2 PM
- Run when user manually triggers it
```

### Step 7: Test Your Agent

**Before deploying:**

1. **Simple Test:** Give it one simple task
   * "Send a test email to [test@company.com](mailto:test@company.com)"
   * Check if email arrives correctly
2. **Complex Test:** Give it a multi-step task
   * "Find a lead, research them, write an email, save to Salesforce"
   * Check each step completed correctly
3. **Edge Case Test:** Test error handling
   * "What happens if email app is down?"
   * "What if data is incomplete?"
4. **Full Simulation:** Run complete workflow on test data
   * 10-20 test scenarios
   * Check for errors or unexpected behavior

### Step 8: Deploy Agent

**When testing passes:**

1. Click **"Deploy"** or **"Publish Agent"**
2. Choose visibility: Private (you only) or Team (shared)
3. Set notifications (email/Slack when agent runs)
4. Configure logging (track all actions)

**Your agent is now LIVE!** 🎉

***

## Agent Builder Features

### 1. Natural Language Description Interface

**How it works:**

* Write in plain English what you want
* Agent Builder interprets your description
* Automatically selects relevant apps and tools
* Suggests rules and workflows

**Example:**

```
Input: "Create an agent that finds leads in tech companies, 
researches them on LinkedIn, and sends personalized emails"

Agent Builder automatically:
- Selects: Google Search, LinkedIn, Email
- Creates workflow: Search → Research → Email
- Sets permissions: Read from web, Write emails
- Suggests trigger: Daily at 9 AM
```

### 2. Visual Workflow Builder

**Components:**

```
┌─────────────────────────────────────────────┐
│  START TRIGGER                              │
│  (When agent runs: Scheduled, Manual, Event)│
└────────────┬────────────────────────────────┘
             │
┌────────────▼────────────────────────────────┐
│  STEP 1: ACTION                             │
│  (Get data, search, create, update)         │
└────────────┬────────────────────────────────┘
             │
┌────────────▼────────────────────────────────┐
│  STEP 2: DECISION                           │
│  (If condition → do X, else do Y)           │
└────────────┬────────────────────────────────┘
             │
┌────────────▼────────────────────────────────┐
│  STEP 3: ACTION                             │
│  (Send email, create task, post to Slack)   │
└────────────┬────────────────────────────────┘
             │
┌────────────▼────────────────────────────────┐
│  STEP 4: LOGGING/NOTIFICATION               │
│  (Report results, log to database)          │
└─────────────────────────────────────────────┘
```

### 3. App Integration Hub

**How it works:**

1. Search for app by name (Gmail, Salesforce, etc.)
2. Click "Connect"
3. Sign in to app account
4. Grant permissions
5. Agent can now use that app

**700+ Supported Apps Including:**

| Category               | Examples                                                |
| :--------------------- | :------------------------------------------------------ |
| **Email**              | Gmail, Outlook, Microsoft Exchange                      |
| **CRM**                | Salesforce, HubSpot, Pipedrive                          |
| **Database**           | Notion, Airtable, Google Sheets                         |
| **Communication**      | Slack, Telegram, Discord                                |
| **Social Media**       | Twitter, LinkedIn, Facebook                             |
| **Productivity**       | Google Calendar, Asana, [Monday.com](http://Monday.com) |
| **Analytics**          | Google Analytics, Mixpanel                              |
| **Payment**            | Stripe, PayPal, Square                                  |
| **Project Management** | Jira, GitHub, Linear                                    |
| **And 690+ more...**   | Check app store in Agent Builder                        |

### 4. Rule Engine

**Define complex logic:**

```
IF-THEN Logic:
├─ IF email from VIP client
│  └─ THEN priority response within 30 min
│
├─ IF deal value > $100,000
│  └─ THEN require manager approval before proceeding
│
├─ IF customer satisfaction score < 3
│  └─ THEN escalate to senior support
│
└─ IF 3 failed attempts
   └─ THEN send human takeover request
```

### 5. Custom Prompts

**You can fine-tune agent behavior:**

```
Base Prompt:
"You are a professional customer support agent.
Always be helpful, empathetic, and solution-focused."

Custom Rules:
- Never promise more than we can deliver
- Always check knowledge base first
- If unsure, ask customer for clarification
- Maximum response length: 500 words

Quality Guidelines:
- Avoid jargon
- Use customer's name when possible
- Provide next steps clearly
- Follow up within 24 hours
```

### 6. Performance Dashboard

**Track agent performance:**

```
📊 Agent Analytics:
├─ Tasks Completed: 245 this week
├─ Success Rate: 94.3%
├─ Average Response Time: 23 minutes
├─ Common Errors: [List top issues]
├─ User Satisfaction: 4.7/5 stars
├─ Cost Savings: $2,340 this month
└─ Recommended Improvements: [Suggestions]
```

### 7. Learning & Improvement

**Agent improves over time:**

* Tracks what works and what doesn't
* Suggests workflow improvements
* Automatically optimizes performance
* You can manually adjust rules
* Agent remembers context from previous runs

### 8. Version Control

**Keep track of changes:**

```
Version History:
├─ v1.0 (Jan 1, 2025) - Initial release
├─ v1.1 (Jan 8, 2025) - Added Slack notifications
├─ v1.2 (Jan 15, 2025) - Improved email template
├─ v2.0 (Jan 22, 2025) - Added LinkedIn integration
└─ Current: v2.0

Rollback: Revert to any previous version instantly
```

***

## Real-World Examples

### Example 1: Sales Outreach Agent

**Description:**

```
"Create an agent that:
- Finds leads in the SaaS industry
- Researches their company and decision makers
- Writes personalized cold emails
- Tracks opens and clicks
- Follows up after 3 days
- Logs all activity in Salesforce"
```

**Apps Used:**

* Google Search (find leads)
* LinkedIn (research companies)
* Gmail (send emails)
* [Hunter.io](http://Hunter.io) (find emails)
* Email tracking tool (track opens)
* Salesforce (log activity)

**Workflow:**

```
1. MORNING (9 AM): Search for leads matching criteria
2. RESEARCH: Get company info from web
3. CREATE: Write personalized email
4. SEND: Email to prospect
5. TRACK: Monitor email opens/clicks
6. FOLLOW-UP (Day 3): Send follow-up if no response
7. LOG: Save interaction to Salesforce
8. REPORT (Weekly): Send summary to sales manager
```

**Result:** Sales team gets 20-30 qualified leads per week, fully researched and tracked, without manual work.

***

### Example 2: Customer Support Agent

**Description:**

```
"Build an agent that:
- Monitors support tickets 24/7
- Reads customer questions
- Searches knowledge base for answers
- Writes helpful responses
- Tags tickets by category
- Escalates complex issues to humans
- Sends satisfaction survey after resolution"
```

**Apps Used:**

* Zendesk/Freshdesk (ticket system)
* Notion (knowledge base)
* Gmail (send responses)
* Slack (notify human team)
* Google Forms (satisfaction survey)

**Workflow:**

```
1. CHECK: Look for new support tickets every hour
2. READ: Parse customer question
3. SEARCH: Find relevant knowledge base articles
4. DECIDE: Can AI handle or needs human?
5. IF SIMPLE: Write response, send ticket, close
6. IF COMPLEX: Summarize issue, send to Slack for human
7. TAG: Label ticket (billing, technical, feature request, etc.)
8. FOLLOW-UP: Send survey 2 hours after resolution
9. TRACK: Log interaction for training
```

**Result:** 70% of tickets resolved automatically within 15 minutes. Humans only handle complex cases.

***

### Example 3: Content Creation Agent

**Description:**

```
"Create an agent that:
- Generates 5 social media posts per week
- Researches trending topics
- Writes posts optimized for each platform
- Schedules posts for best engagement times
- Monitors comments and engagement
- Reports weekly performance"
```

**Apps Used:**

* Google Trends (trending topics)
* ChatGPT/Claude (write content)
* Buffer/Hootsuite (schedule posts)
* Twitter, LinkedIn, Facebook (publish)
* Google Analytics (track engagement)
* Slack (send reports)

**Workflow:**

```
MONDAY:
1. RESEARCH: What's trending this week?
2. BRAINSTORM: Generate 10 post ideas
3. WRITE: Create posts for each platform
4. SCHEDULE: Queue for best times
   - Twitter: 9 AM, 3 PM, 8 PM
   - LinkedIn: 8 AM, 12 PM
   - Facebook: 10 AM, 5 PM

TUESDAY-FRIDAY:
1. MONITOR: Track comments/likes
2. ENGAGE: Respond to comments
3. TRACK: Update engagement metrics

FRIDAY EVENING:
1. REPORT: Weekly performance summary to Slack
2. ANALYSIS: What worked? What didn't?
3. RECOMMEND: Ideas for next week
```

**Result:** Consistent, high-quality social media presence with 24/7 monitoring, no manual effort after setup.

***

### Example 4: HR Interview Coordinator Agent

**Description:**

```
"Build an agent that:
- Reviews resumes against job requirements
- Scores candidates automatically
- Sends interview invitations to qualified candidates
- Schedules interviews in calendar
- Sends interview prep materials
- Collects feedback from interviewers
- Updates applicant tracking system"
```

**Apps Used:**

* Email (receive resumes)
* Google Drive (store resumes)
* Job board API (job requirements)
* Google Calendar (schedule interviews)
* Gmail (send invitations)
* Slack (notify hiring team)
* Typeform (collect feedback)
* ATS (Lever, Greenhouse, etc.)

**Workflow:**

```
STAGE 1: RESUME SCREENING
1. RECEIVE: New resume arrives
2. EXTRACT: Pull candidate info
3. SCORE: Compare to job requirements
4. DECIDE: Pass to interview or reject?

STAGE 2: QUALIFIED CANDIDATES
1. DRAFT: Write personalized email
2. ATTACH: Prepare interview materials
3. SEND: Invitation to candidate
4. SUGGEST: 5 available interview slots
5. UPDATE: Add to calendar hold

STAGE 3: INTERVIEW SCHEDULING
1. WAIT: For candidate response
2. MATCH: Calendar availability
3. CONFIRM: Book interview slot
4. NOTIFY: Hiring team in Slack
5. SEND: Prep materials to candidate

STAGE 4: POST-INTERVIEW
1. COLLECT: Interviewer feedback
2. SCORE: Combine feedback
3. UPDATE: ATS with results
4. NOTIFY: HR team
5. NEXT STEP: Schedule next round or send rejection
```

**Result:** HR team processes 50+ applications per week with zero manual effort. Best candidates are identified and scheduled in 24 hours.

***

### Example 5: Data Analyst Agent

**Description:**

```
"Create an agent that:
- Collects data from multiple sources
- Cleans and processes data
- Creates visualizations and reports
- Identifies trends and anomalies
- Sends weekly summary reports
- Flags metrics that need attention"
```

**Apps Used:**

* Google Analytics
* Salesforce
* Google Sheets
* DataBox (visualization)
* Slack
* Email

**Workflow:**

```
WEEKLY PROCESS:

1. COLLECT DATA
   - Website traffic from Google Analytics
   - Sales numbers from Salesforce
   - Customer satisfaction scores
   - Social media metrics

2. PROCESS
   - Combine all data sources
   - Remove duplicates/errors
   - Normalize formats
   - Calculate year-over-year changes

3. ANALYZE
   - Identify top performers
   - Find problem areas
   - Calculate trends
   - Spot anomalies

4. VISUALIZE
   - Create charts and graphs
   - Build interactive dashboard
   - Highlight key metrics

5. REPORT
   - Send weekly email summary
   - Post key metrics to Slack
   - Alert team to issues
   - Recommend actions
```

**Result:** Managers have complete business overview every Monday morning, automatically generated.

***

## Best Practices

### 1. Start Simple, Then Expand

**Don't:**

```
"Create an agent that does everything our company needs:
handle sales, support, marketing, HR, accounting, etc."
```

**Do:**

```
"Create a sales agent that finds leads and sends emails"

(Then create separate agents for support, marketing, etc.)
```

### 2. Be Specific in Descriptions

**Vague:**

```
"Create a sales agent"
```

**Clear:**

```
"Create an agent that:
- Finds tech companies with 10-50 employees in US
- Researches their tech stack on web
- Finds decision makers on LinkedIn
- Writes 3-4 sentence cold emails about our X product
- Sends Monday-Friday only
- Tracks opens and clicks
- Follows up after 3 days if no response"
```

### 3. Give Clear Rules

**Without rules:**

* Agent might make mistakes
* Inconsistent quality
* Unexpected behavior

**With rules:**

```
ALWAYS:
- Check facts before writing
- Use professional tone
- Include call-to-action
- Add signature with contact info

NEVER:
- Make promises we can't keep
- Use spam language
- Send to competitors or blacklisted domains
- Send on weekends or holidays

IF [condition] THEN [action]:
- IF customer angry THEN escalate to human
- IF deal > $50k THEN get manager approval
- IF no response 3x THEN mark as low priority
```

### 4. Test with Real Data First

**Best approach:**

1. Create agent
2. Test with 5-10 real scenarios
3. Review outputs manually
4. Refine rules based on results
5. Then deploy fully

**Don't:**

* Deploy immediately without testing
* Test only with made-up examples
* Assume it will work correctly

### 5. Monitor Performance

**Track metrics:**

* Success rate (% of tasks completed)
* Error rate (% of failures)
* Quality (user satisfaction)
* Speed (time to complete)
* Cost (credits used)

**Adjust accordingly:**

* If success \< 90%, improve rules
* If quality scores low, refine prompts
* If too slow, optimize workflow
* If costs high, reduce scope

### 6. Document Everything

**Create a simple guide:**

```
AGENT: Sales Outreach
PURPOSE: Find and email qualified leads
CREATED: Jan 1, 2025
OWNER: Sarah (sales@company.com)

WHAT IT DOES:
1. Searches for leads daily
2. Researches company info
3. Writes and sends personalized emails
4. Follows up after 3 days

HOW TO USE:
- It runs automatically every weekday at 9 AM
- No action needed from you
- Check results in Salesforce CRM

CONTACT IF:
- It stops working or makes errors
- You want to change rules
- You want to add new features
```

### 7. Iterate Based on Feedback

**Weekly improvement cycle:**

```
WEEK 1: Deploy agent
WEEK 2: Review outputs, gather feedback
WEEK 3: Identify what's wrong
WEEK 4: Update rules/prompts
WEEK 5: Test changes
WEEK 6: Deploy improved version
Repeat...
```

***

## Troubleshooting

### Problem: Agent is making errors

**Solution:**

1. Check the rules - are they clear enough?
2. Review agent's logic - is it following instructions?
3. Refine the description - add more context
4. Add specific rules to prevent errors

```
IF [error type] THEN [prevention]
IF agent writes factually wrong info THEN "Always verify facts in knowledge base before writing"
IF agent uses wrong tone THEN "Always write in professional, not friendly tone"
```

### Problem: Agent is too slow

**Solution:**

1. Reduce number of steps in workflow
2. Remove unnecessary data processing
3. Pre-filter data before sending to agent
4. Run agent less frequently if possible

### Problem: Agent is using too many credits

**Solution:**

1. Reduce frequency (run daily instead of hourly)
2. Process fewer items per run
3. Use simpler AI models for simple tasks
4. Batch similar tasks together

### Problem: Agent keeps missing cases

**Solution:**

1. Add catch-all rules for edge cases
2. Improve trigger conditions
3. Lower thresholds for escalation
4. Add logging to identify patterns

### Problem: Certain apps won't connect

**Solution:**

1. Check that you have admin access to the app
2. Verify the app supports API connections
3. Try reconnecting: remove and re-add the app
4. Contact NeuroAI support if error persists

### Problem: Agent is doing things you didn't ask for

**Solution:**

1. Add "NEVER" rules to restrict behavior
2. Be more specific in your description
3. Add validation steps before actions
4. Reduce agent's autonomy (require human approval for some actions)

***

## FAQs

### Can I change an agent after it's deployed?

Yes! You can:

* Update the description
* Modify rules
* Add/remove apps
* Change the trigger schedule
* Version control changes

Changes take effect immediately (or on next run, depending on type).

### What if I need multiple similar agents?

You can:

* Duplicate an existing agent
* Modify the duplicate
* Deploy as separate agent

This is faster than creating from scratch.

### Can agents work together?

Yes! You can chain agents:

* Agent 1 finds leads
* Agent 2 researches them
* Agent 3 sends emails
* Agent 4 tracks results
* Agent 5 reports weekly

Each agent triggers the next when complete.

### How much does it cost?

Agent Builder is free with any NeuroAI plan.

You only pay for:

* Credits used for AI processing
* API calls to connected apps (if app charges)

Typical costs:

* Small agent: 50-100 credits/month
* Medium agent: 200-500 credits/month
* Large agent: 500-2,000 credits/month

### Is my data safe?

Yes!

* All data encrypted in transit and at rest
* Apps are only accessed with your permission
* You can revoke access anytime
* No data shared with other users
* GDPR compliant

### Can agents handle errors?

Yes!

* Agent Builder creates error handling automatically
* If app fails, agent can retry or use alternate path
* Can escalate to human if unrecoverable
* Logs all errors for debugging

### What happens if the internet goes down?

If YOUR internet: Agent continues running (it's cloud-based) If NEURO's internet: Agent pauses, resumes when online

All actions are logged so you won't lose work.

### Can I use Agent Builder on mobile?

Yes! The dashboard works on:

* Mobile web (recommended)
* Coming soon: Native mobile app

Best experience is desktop for first-time setup, but you can monitor agents on phone.

### How many agents can I create?

Unlimited! You can create as many custom agents as you need. No per-agent fees.

### Can my team see my agents?

You choose!

* **Private:** Only you can see/use
* **Team:** All team members can see/use
* **Organization:** Everyone in company can see/use

You control permissions and sharing.

***

## Getting Started Now

### Quick Start (5 minutes)

1. **Log into NeuroAI** → Dashboard
2. **Click** "Create Custom Agent"
3. **Describe** what you want (1-2 paragraphs)
4. **Select** apps from recommendations
5. **Click** "Deploy"

Your agent is now running!

### Learn More

* **Tutorials:** Visit NeuroAI Help Center
* **Examples:** Browse Agent Templates
* **Community:** Join NeuroAI Slack for tips and tricks
* **Support:** Email [info@patternautomation.com](mailto:info@patternautomation.com)

### Popular Starting Points

* **Sales Agent:** Get qualified leads automatically
* **Support Agent:** Handle customer questions 24/7
* **Content Agent:** Create social posts weekly
* **HR Agent:** Screen resumes and schedule interviews
* **Data Agent:** Generate weekly business reports