AI Enabled Professionals

Learning Outcomes
1. Understand agents, the brains behind the operation Learn to define a CrewAI agent using role, goal, backstory, LLM, and tools so it knows what job it owns.
2. Write tasks that tell agents what to deliver Learn to build a task with description, agent, expected_output, and context so the result is predictable.
3. Assemble a crew that runs agents and tasks Learn to use the crew as the container that brings your agents and tasks together and actually runs them.
4. Run crews sequentially, concurrently, or hierarchically Learn to pick the right processing mode and see how design patterns map to how your crew executes work.
5. Build a working multi-agent crew end to end Learn to connect multiple agents, tasks, and a crew into one run where agents hand off work to each other.
Why This Topic Matters
One agent hits a ceiling fast. Real work has steps, handoffs, and specialists, and a single prompt cannot carry all of it. CrewAI gives you agents, tasks, and a crew that runs them in sequential, concurrent, or hierarchical order. You leave able to break a messy process into roles a system can run, which is how AI moves from answering questions to doing the job.

Introducing 🧪Lab Sessions. This session will be structured as follows:

First 20 minutes
- Demonstrating the agent we are going to build
- Providing you with the prompts and step-by-step workflow. Only Available To Attendees.

Remaining Session:
- You will build, and we will be on standby to help when you get stuck.
- You will leave with having built the agent.

What You Will Need For This Lab:
- Gmail Account
- ChatGPT Account (Ideally a paid tier subscription)
- Watch The ChatGPT Session In The Resources If You Did Not Attend

Goal: You will learn ChatGPT Work and add a real working use case to your portfolio.

PLEASE NOTE: THERE ARE NO RECORDINGS FOR THIS SESSION.

Learning Outcomes
1. Connect Gmail and Sheets to ChatGPT Work
Learn to connect Gmail and Google Sheets to ChatGPT Work so your agent can read, draft, and log.
2. Build an agent that writes your draft emails
Learn to write the agent instructions that turn a contact list into personalized Gmail drafts.
3. Label and separate drafts inside Gmail
Learn to have the agent label every draft in Gmail so outreach stays separate from your real inbox.
4. Schedule daily sends that stay out of spam
Learn to send drafts on a daily schedule in small batches so Gmail does not flag your outreach.
5. Update your CRM in Google Sheets automatically
Learn to log every send, reply, and follow up date back into your Google Sheets CRM without typing.

Why This Topic Matters
Outreach is the part of your job that never scales. Email one at a time and you run out of hours. An agent that drafts in Gmail, labels the queue, sends daily, and logs to Sheets gives you a working outreach system you own. You leave able to wire an agent into the apps you already use every day, which is the skill that makes AI actually pay off at work.

A working interface does not automatically mean an application is secure. New builders often focus on pages and features while assuming that signing in protects all the data. But app access and data access are different. Without appropriate RLS rules, users may be able to view or modify records they should never access.

What you'll learn

⚡Protect User Data with RLS
Learn how Row-Level Security controls which records users can create, read, update, and delete inside your Base44 app.
⚡Configure and Test Data Access
Learn to apply a simple ownership-based permission rule and verify the result from different users’ perspectives.
⚡Fix a Bug with a Focused Prompt
Learn to describe the affected user, action, expected behavior, and actual result so Base44 can make a controlled repair
⚡Invite a Collaborator Safely
Invite someone to help build your app while understanding the difference between collaborators, app admins, & users