How I AI: Hiten Shah's Advanced ChatGPT Workflows for a Personal Coach, Sales Playbooks, and Managing Up

In this episode, I was so excited to sit down with Hiten Shah, a serial founder who has been building B2B SaaS for over 20 years. He's the mind behind companies like Crazy Egg and KISSmetrics, and he’s one of the most advanced ChatGPT users I’ve ever met.

Hiten has moved way beyond simple question-and-answer prompts. He uses ChatGPT Projects to create persistent, context-aware AI assistants that act as personal coaches, sales strategists, and communication guides. His whole approach is built on a simple idea: to get great outputs, you have to provide great inputs. That means doing the work up front to give the AI context, frameworks, and—most importantly—clear examples of what 'great' looks like.

In our conversation, Hiten shared three workflows that you can copy and use right away. We'll walk through how to create an AI version of your boss for practicing difficult conversations, how to build a 'Personal OS' for AI-driven self-coaching, and how to turn any sales methodology into a tool that generates custom scripts on demand. These are real, systematic ways to weave AI into your daily work to become more effective, self-aware, and persuasive. Let's get into it.

Workflow 1: Replicating Your Boss with a "What Would Morgan Do" Project

One of the trickiest things to navigate at work is managing up. How do you pitch a risky idea? Ask for help? Or just frame a conversation so it lands with your manager? Hiten showed me a fascinating workflow for creating an AI ‘clone’ of his boss, Morgan, to rehearse these exact situations.

This is such a useful idea for anyone on a team. Instead of guessing how your manager might react, you can run a simulation with an AI that’s been trained on their specific communication style and priorities.

Step 1: Create a New ChatGPT Project

First, Hiten starts by creating a new Project in ChatGPT. This feature is what makes the whole thing work, because it lets you upload files and set custom instructions that stay in place for every conversation in that project. It creates a dedicated space that’s already loaded with context.

Step 2: Gather and Upload Contextual Files

The real secret to this workflow is the source material. Hiten gathered two key documents:

• Morgan's Operating Manual: A document Morgan wrote himself explaining how he works, his expectations, and his communication preferences.
• Favorite Articles: Articles Morgan had shared that reflect his thinking and values, such as one titled "Job is Communication."

He uploads these files directly into the project, giving the AI a rich, primary source of information about his boss from the get-go.

Step 3: Generate Project Instructions with a Meta-Prompt

Instead of writing the project instructions himself, Hiten gets ChatGPT to create them. He starts a new chat, uploads the files, and asks the AI to define its own role. This meta-prompting trick is a very clever shortcut.

I'm creating a project with these files and I want to be able to converse with it... and simulate the type of feedback or advice that my boss would give me. Can you create the instructions for that project?

ChatGPT then produces a detailed set of instructions outlining its persona (Morgan), its objective (to provide feedback based on his known principles), and how it should use the provided documents. Hiten just copies and pastes these directly into the project's instruction settings.

Step 4: Test the Project with a Real-World Scenario

With the project set up, it was time to put it to the test. Hiten posed a classic workplace challenge:

I wanna pitch Morgan the craziest products idea I can think of. What is the best way to pitch it to him so we can go after it.

The result was uncannily accurate. The AI didn't give generic advice; it gave Morgan's advice. It suggested leading with a specific vision, anticipating his questions about data, and framing the pitch in a way that aligns with his documented values. Hiten’s reaction said it all: "Yeah, that sounds like him." This is a fantastic way to prepare for any high-stakes conversation and seriously improve your chances of success.

Workflow 2: Building a "Personal OS" for AI-Powered Self-Coaching

After showing us how to model his boss, Hiten shared an even more personal workflow: creating an AI coach for himself. This "Personal OS" is designed to give him advice, help him reframe challenges, and navigate interpersonal dynamics by understanding his own personality and communication style.

Think of it as having an executive coach on-call 24/7. It can help you make sense of your own emotional reactions, handle conflict better, and build stronger relationships with colleagues by helping you see things from their side.

Step 1: Gather Your Personal Framework Data

Just like the 'Morgan' project, this one is built on a foundation of good context. Hiten uploaded documents and information about his own personality based on several well-known frameworks:

• Enneagram: Hiten is a Type 9.
• Human Design: He has a 1/3 profile.
• Myers-Briggs: His type information.
• DISC: His assessment results.
• Voice and Tone Guide: A personal branding document created for him in the past.

Step 2: Combine Your Personal OS with External Context

This is where the approach gets really interesting. Hiten showed how to use his Personal OS to better understand his dynamic with his boss, Morgan. He uploaded Morgan's operating manual into his own coaching project and prompted it:

This manual is from my boss Morgan. I'm newly reporting to him and would love any and all advice on how best to work with him considering my own personality, etc. By the way, he's an Enneagram type five.

The AI provided a really sharp analysis, contrasting his Enneagram 9 tendencies (peacemaking, consensus-seeking) with Morgan's Enneagram 5 traits (analytical, detached, knowledge-seeking). It offered specific strategies for communication, like leading with data-driven arguments and respecting Morgan's need for independent thought. It even suggested a weekly check-in format designed to suit both their styles.

Step 3: Use it for Real-Time Emotional Coaching

The Personal OS can also be a tool for managing your own state of mind. Hiten gave an example of a prompt he might use when feeling frustrated:

I'm really mad right now about someone trying to get control over one of the projects I'm working on because they are trying to steal the cookie.

The AI's response was empathetic and insightful. It started by validating his feelings ("This reaction makes perfect sense given your profile") and then reframed the issue. It wasn't just a threat to his control, but a threat to the integrity of his work—a distinction that really resonated with him. This kind of perspective shift is exactly what a great coach does.

Workflow 3: Operationalizing a Sales Framework with ChatGPT

For the final workflow, Hiten showed how to take a static corporate framework and turn it into a dynamic, interactive tool. Many of us work at companies that have specific methodologies—for sales, product development, or marketing—that usually just sit in dense PDFs and PowerPoints. Hiten used the Winning by Design sales framework as an example to show how to make it useful.

Step 1: Source and Upload the Framework Documents

First, Hiten became a domain expert—or rather, he made ChatGPT one. He used Google to find all the publicly available PDFs on the Winning by Design framework and its different processes (like the SPICE model for discovery calls). He then created a new ChatGPT project and uploaded all of those documents.

Step 2: Generate a Customized Discovery Guide

With the AI now an expert on the framework, Hiten tasked it with creating a practical sales asset. He asked it to apply the framework to my product, ChatPRD:

Can you create a SPICE discovery guide for ChatPRD?

The AI instantly generated a detailed discovery call script. Instead of just a list of questions, it produced a structured guide organized around the SPICE framework: Situation, Pain, Impact, and Critical Event. The questions were specific and designed to uncover customer needs in a way that aligns perfectly with the methodology.

Step 3: Enrich with Deeper, Product-Specific Context

The initial script was good, but it was based on limited public knowledge of ChatPRD. To make it even better, Hiten ran a separate 'deep research' prompt on ChatPRD to gather more details. He then copied this new context and fed it back into the sales project with a simple instruction:

Now here is context on ChatPRD... please improve the deliverable above using this new context.

This two-step process—first teaching the AI the framework, then giving it the specific subject matter—is what lets you create such highly tailored and effective outputs. It immediately refined the discovery script with more nuanced questions that showed a deeper understanding of the product and its target audience.

Conclusion: Build Your Own Specialized AI Assistants

What I love about Hiten’s approach is that it's systematic and you can use it for almost anything. The common thread through all three workflows is the idea of combining structured, external knowledge (manuals, personality frameworks, sales playbooks) with your specific, personal context inside a dedicated ChatGPT Project. This is how you elevate the AI from a general-purpose tool to a set of specialized assistants that are experts in your world.

A big takeaway from Hiten is to be deliberate. Don't rush to automation. First, get comfortable with the manual process of prompting and refining. Figure out what inputs generate the best outputs. As he says, you need to "show it what great looks like." Once you've done that, you can build very effective, repeatable systems that will truly change how you work.

I encourage you to try one of these workflows this week. Make a clone of your boss, build your personal OS, or bring a framework you use at work to life. You have all the tools you need to build your own custom AI team.

Thanks to our Sponsors

A huge thank you to our sponsors for supporting the show!

• Paragon—Ship every SaaS integration your customers want
• Notion—The best AI tools for work

Episode Links

• Watch on YouTube
• Listen on Spotify
• Listen on Apple Podcasts