If you are still using AI just to “write an email” or “fix this code,” you are driving a Ferrari in first gear.
The real power of artificial intelligence right now isn’t in simple Q&A; it is in App Smashing—taking the output of one AI tool and feeding it into another to create something a single model couldn’t do alone.
I have been testing some deeper workflows recently, and I found three techniques that genuinely feel like “secrets” because almost nobody is using them. Here is how to turn your AI tools into a personal university, a financial analyst, and a research lab.
Contents
1. The “Auto-University” Workflow
Tools: Perplexity + NotebookLM
We have all tried to learn a new skill (like Python or Marketing) by asking ChatGPT, only to get a wall of text that is exhausting to read. Here is a better way to design your own curriculum.
The Workflow:
-
Curate with Perplexity: Instead of asking for a lesson, ask Perplexity for a list of resources.
-
Prompt: “I want to learn Python. Give me the top 20 URLs for tutorials and documentation. Give me the URLs only.”
-
-
Digest with NotebookLM: Take those URLs and paste them into Google’s NotebookLM. This is a free tool designed to “ground” AI in specific sources.
-
Generate a Lecture: Once the sources are added, click the “Audio Overview” button.
The Result: NotebookLM will generate a stunningly realistic “podcast” where two AI hosts discuss the material you uploaded. You can listen to your research while you drive, effectively turning any topic into a bespoke university lecture. You can even ask it to design a step-by-step curriculum based only on those high-quality links.
2. The “Instant Analyst” Workflow
Tools: Perplexity (Finance Mode) + Google Gemini (Canvas)
Imagine your boss asks you to analyze a competitor’s stock performance and present it in a visual dashboard. Usually, this takes hours of digging through SEC filings and wrestling with Excel. You can do it in minutes.
The Workflow:
-
Deep Search: Go to Perplexity and toggle on “Finance” mode (this focuses the AI on financial data and stock markets).
-
Prompt: “Analyze the SEC filings for Starbucks (SBUX). Look for growth, profitability risks, and red flags.”
-
-
Visualize with Canvas: Copy the detailed text report from Perplexity. Open Google Gemini and ensure “Canvas” mode is active. Paste the text.
-
Transform: Ask Gemini to “Turn this data into an interactive website” or “Create an infographic.”
The Result: Gemini Canvas won’t just rewrite the text; it will actually code a functional HTML webpage or generate a visual infographic chart right before your eyes. You go from raw data to a boardroom-ready presentation in about 90 seconds.
3. The “Permanent Expert” Workflow
Tools: ChatGPT (Deep Research + Projects)
One of the biggest frustrations with AI is that it “forgets” who it is supposed to be. You tell it to “act as a copywriting expert,” but three messages later, it sounds generic again. The solution is to build a Project.
The Workflow:
-
Conduct Deep Research: First, ask ChatGPT to research a topic exhaustively.
-
Prompt: “Do deep research on the psychology of high-converting copywriting. Write a comprehensive guide.”
-
-
Create the Artifact: Take that long-form research and save it as a PDF or a clean document.
-
Build the Project: In ChatGPT, go to “Projects” and create a new one called “Copywriting Master.” Upload your research PDF here.
The Result: Now, whenever you enter this “Project” folder, ChatGPT references that specific file first. You don’t need to keep prompting it to “act like an expert”—it is the expert because you gave it a textbook to reference. You have effectively built a custom employee trained specifically on your data.
Final Thoughts
The gap between “average” AI users and “power” users is widening. The average user types a prompt and hopes for the best. The power user understands that AI tools are modular—Perplexity is for finding, NotebookLM is for learning, and Gemini is for visualizing.
Start connecting these dots, and you will find you can accomplish week-long projects in a single afternoon.
