How to Build Your Own AI Mentor Using ChatGPT, Deep Research, and Audio Transcription

Artificial intelligence is rapidly transforming how we learn, think, and access knowledge. Instead of simply reading books or watching videos, you can now build a personalized AI mentor trained on the teachings of your favorite thinkers, coaches, or authors. Imagine having an AI version of a spiritual teacher, business coach, or expert available anytime to answer questions and guide your decisions.

In a recent tutorial, Marco from Nari Media demonstrated how anyone can create a custom AI mentor using ChatGPT’s Custom GPT builder, a curated knowledge base, and transcription tools to convert audio content into text. The process is surprisingly accessible and opens the door to building deeply personalized AI coaching systems.

Let’s walk through the core steps.

Step 1: Create a Custom GPT in ChatGPT

The foundation of your AI mentor starts with ChatGPT’s Custom GPT builder.

To begin:

Open ChatGPT.
Click your profile icon in the top-right corner.
Select “My GPTs.”
Click “Create a GPT.”

This opens a configuration panel where you can customize your AI assistant. You will need to define several key components:

Name – What the mentor will be called
Description – What the AI does and how it helps users
Instructions – Detailed behavior rules and tone
Conversation starters – Questions users can ask immediately
Knowledge base files – PDFs, documents, or transcripts the AI can reference

While ChatGPT can automatically generate a template, manually configuring the assistant gives you greater control and better results.

Step 2: Use ChatGPT to Write the Perfect Instructions

Instead of writing instructions yourself, you can ask ChatGPT to design them for you.

For example, you might prompt ChatGPT with something like:

“Help me build a custom GPT that acts as a mentor based on the teachings of a specific author. It should reference uploaded PDFs and transcripts, respond in the author’s style, and provide thoughtful coaching.”

ChatGPT can then generate:

A name for the AI mentor
A detailed description
Expanded instructions
Suggested conversation starters

These instructions shape how the AI behaves. For example, if the mentor is based on a spiritual teacher, the instructions might specify that the AI should:

Speak calmly and thoughtfully
Use reflective language
Avoid sounding like a generic chatbot
Focus on presence, awareness, and personal insight

The more detailed your instructions, the more authentic your AI mentor will feel.

Step 3: Build a Powerful Knowledge Base

The knowledge base is the most important part of your AI mentor.

As the principle goes:

Better input leads to better output.

You can upload many types of content:

PDF books
Lecture transcripts
Research reports
Articles
Notes
Text documents
Course materials

For example, if you wanted to build an AI mentor based on a philosopher or coach, you could upload:

Their books
Interviews
seminar transcripts
lecture notes
essays

The AI will then use this information as a reference when answering questions.

You can even collect YouTube transcripts and add them to your knowledge library.

The larger and more detailed your knowledge base becomes, the more insightful your AI mentor will be.

Step 4: Use Multiple AI Models for Deep Research

To expand the knowledge base further, you can gather research using several AI systems.

Marco demonstrates running the same research prompt across multiple platforms, including:

ChatGPT
Google Gemini
Grok

Each AI searches different sources and compiles information in slightly different ways. By combining the outputs from all three, you create a much richer knowledge database.

The process looks like this:

Write a deep research prompt asking for detailed analysis of a subject or teacher.
Run the prompt in multiple AI systems.
Export the research results.
Combine them into a single document.

This compiled document becomes a powerful reference file you can upload into your Custom GPT.

Step 5: Compile Research into a Master Document

Once research from multiple sources is gathered, place everything into a single master document.

You can do this with:

Google Docs
Microsoft Word
Any text editor

The document may contain:

summaries
philosophical concepts
common questions
teaching styles
core principles
terminology

This file effectively becomes the brain of your AI mentor.

After saving the document, simply upload it into the Custom GPT knowledge section.

Step 6: Convert Audio Books to Text with Whisper

Many people own large collections of audio courses or MP3 audiobooks. These can also be turned into knowledge sources.

To do this, you can use Whisper, an AI transcription tool.

The process requires three main components:

Python
FFmpeg
Whisper AI

After installing these tools, you can run a command in your computer’s terminal that converts audio files into text transcripts.

The command typically looks something like this:

python -m whisper filename.mp3

Once the transcription finishes, Whisper automatically generates a TXT file containing the entire transcript.

This text file can then be uploaded into your Custom GPT knowledge base.

By converting multiple audiobooks into text, you can train your AI mentor on thousands of pages of information.

Step 7: Add Conversation Starters

To make the AI easier to use, add conversation starters such as:

“How can I become more present in daily life?”
“What does it mean to live in the present moment?”
“How do I handle anxiety or stress?”
“How can I quiet my mind?”

These prompts help users immediately interact with the mentor.

They also guide the AI toward the types of conversations it was designed for.

Step 8: Customize the AI Avatar

For a fun finishing touch, you can generate a visual avatar for your AI mentor.

Using an image generator, you might create a character that represents the mentor’s personality — for example:

A futuristic enlightened sage
A cybernetic philosopher
A digital monk
A holographic guide

This image becomes the icon for the Custom GPT.

Step 9: Publish or Keep Private

Once the AI mentor is complete, you can choose how it will be shared.

Options include:

Private use only
Shareable via link
Published in the GPT store

Publishing allows others to interact with your AI mentor as well.

The Real Power of AI Mentors

The concept of AI mentors is powerful because it turns static information into interactive wisdom.

Instead of searching through books, users can ask questions like:

“What should I do when I feel overwhelmed?”
“How do I stay focused on my goals?”
“What does success really mean?”

The AI synthesizes information from its knowledge base and delivers thoughtful responses instantly.

Over time, as you add more material, the mentor becomes increasingly intelligent.

The Future of Personalized AI Learning

We are entering an era where everyone can build personalized digital teachers.

Students can create AI tutors trained on textbooks.
Entrepreneurs can create AI business coaches.
Spiritual seekers can build AI philosophy mentors.
Investors can build AI market analysts.

Instead of passively consuming information, we can now interact with knowledge directly.

And as AI tools continue improving, these mentors will become more accurate, more contextual, and more powerful.

The ultimate result is a new way of learning — where wisdom, expertise, and guidance are available instantly, anytime you need it.

How to Build Your Own AI Mentor Using ChatGPT, Deep Research, and Audio Transcription