Big Idea
AI is most useful when it is treated as a practical working partner, not a novelty, a magic trick, or a replacement for judgment.
This lesson sets the frame for the AI portion of the course. The point is not to claim that one tool is perfect or that every broker needs to become a technical expert. The point is that AI is now useful enough to become part of daily work, especially when it is connected to the knowledge and systems developed earlier in the course.
Why This Lesson Matters
There are more AI presentations now than there were a few years ago. Agents are seeing AI sessions at industry meetings, conferences, and webinars. Some of those sessions are useful, but many still stay at the level of general excitement or simple writing examples.
This lesson explains the perspective behind the rest of Part 3. It comes from a combination of formal AI education and heavy real-world use. That combination matters because AI is not best understood by reading about it once. It is best understood by using it repeatedly until it becomes part of how you work.
AI Education Background
Eric’s AI background includes multiple formal programs. The first was a nine-month artificial intelligence and machine learning program through UT Austin and Great Learning, completed in early 2024. That program included self-paced lessons, quizzes, tests, weekend mentor sessions, and Python notebooks.
The value of that program was not that it turned him into a full-time coder. The value was that it made the underlying technology less mysterious. It gave him a better sense of how these tools work and why they can understand language, patterns, and relationships between ideas.
He also attended a three-day AI immersion program in Austin in 2024, where he built his first RAG system. Later, he completed a three-month program focused on large language models, retrieval augmented generation, and agentic AI. Those topics are the same practical layers this part of the course will introduce.
Important distinction: formal education helps, but the real learning curve comes from repeated use. AI starts to make sense when you use it in real workflows, not just when you hear someone describe it.
Usage Matters More
The more important credential is usage. Eric has used AI almost daily since shortly after ChatGPT was released to the public in late 2022. That consistent use is what makes the lessons practical.
AI has been used in business workflows, personal workflows, writing, research, brainstorming, class development, website work, tool development, and ongoing experimentation. That does not mean every answer is perfect or every output should be trusted without review. It means the lessons come from repeated interaction with the tools.
That matters because AI has a learning curve. You learn what to ask, when to give more context, when to push back, when to revise, and when to stop and use your own judgment.
Pick One Tool and Learn It Well
There are several strong AI tools available. Claude is excellent for some tasks, especially longer documents and long-form writing. Other tools may be better for research, coding, search, or specific workflows.
But for someone still learning, the best advice is usually to pick one tool and learn it well before jumping between several tools. Bouncing back and forth too early can slow the learning process because you are constantly comparing tools instead of getting better at using one.
The tool used in this course is ChatGPT. The larger lesson is not that every person must use ChatGPT forever. The lesson is that mastery comes from consistent use, not from constantly switching platforms.
Assume It Can Help
A useful mindset is to assume the tool can probably help, then have a conversation with it about how. That is the same mindset Eric used when building the Stella quoting spreadsheet years ago. He did not know every Excel formula, but he assumed the functionality probably existed and kept searching until he found a way.
AI works well with that same approach. Instead of starting with, “Can it do this?” start with, “Here is what I am trying to accomplish. How can you help me do it?”
That shift matters because most people underestimate what AI can do. They use it for simple writing help when it can also assist with structuring ideas, building workflows, creating learning materials, extracting information, reusing assets, and developing tools.
What AI Does
At a basic level, AI tools help you think faster, write faster, and process information faster. GPT stands for generative pre-trained transformer. The word generative matters because these tools create new text, not merely return existing web pages like traditional search.
Large language models are trained on huge amounts of text. They learn patterns in language and relationships between words, sentences, and ideas. That is why they can understand a rough explanation, a rambling prompt, or an incomplete thought and still produce a useful response.
You do not need to understand every technical detail to use these tools well. It is fine to be curious about how they work, but the more important question for an agency is practical: how can this help with the work we actually do every day?
AI Is the Multiplier
Everything in Part 2 was about leverage: reducing work, reusing what you have already done, building systems, creating assets, using tools, and designing workflows instead of constantly reacting.
AI fits on top of that. It is not the starting point. It is the multiplier. Multipliers only work when there is something real to multiply.
That is why the course spent two parts building the foundation before moving into AI. Public chatbots can already help with emails, grammar, blog posts, summaries, and basic drafts. That is useful, but it is only the beginning. AI becomes much more valuable when it can work with your information, your processes, your preferences, and your agency knowledge.
AI is not the starting point. It is the multiplier.
A Note About HIPAA
When using AI in the benefits world, be careful with protected health information. A practical rule is not to put HIPAA-protected information into a public AI chat. In most cases, there is no need to do that.
This is one reason to separate client decision-making information from claims, medical, and service-history information. A CRM may contain secure records about claims, dates of birth, medical issues, or service problems. That information should not simply be copied into an AI tool for analysis.
For AI purposes, the more useful information is usually non-HIPAA decision context: how the client makes decisions, what plans they prefer, what strategies they rejected, what constraints matter, and how the agency usually positions options. That kind of structure can make AI more useful without unnecessarily exposing protected information.
Apply It
Choose one AI tool to focus on for the next few weeks. Use it regularly for real tasks instead of only testing it occasionally.
As you use it, pay attention to what improves the output: better context, clearer goals, examples, follow-up questions, and your own judgment. That repeated use is what turns AI from an interesting tool into a practical workflow partner.