Prompt Engineering for TPMs

Steal 3 of my TPM Prompts

Together with

(6 minute read)

👋 Hey TPM Craftsman, let’s get crafting.

This newsletter edition is focused on the Technical Versatility

What’s inside?

👨‍🏫 Learn: Prompt Engineering for TPMs (steal my prompts!)
đź“š Resource: The TPM Craft Advisor GPT is my go-to AI agent for all-things TPM.
đź‘‹ Sponsor: The Rundown AI. A newsletter that I personally subscribe to for AI news.

Let’s get to it! 👇

Learn

👨‍🏫 Prompt Engineering for TPMs: time to adapt and multiply your impact

In the world of artificial intelligence and machine learning, prompt engineering is quietly loudly stealing the spotlight.

Why? Its the most user-friendly, low-barrier-to-entry way for the general public to interact with AI in a meaningful way.

TPM’s will soon lean on prompt engineering as a key amplifier of all the core skills.

Those who don’t adapt will fall behind.

Those who do adapt will multiply their impact.

Why should you care?

Why should you care about improving your prompt engineering skills as a TPM? 3 reasons:

  1. Learn more quickly

  2. Boost your productivity

  3. Multiply your impact

These benefits build on each other. The more capable you are at learning, the more productive you’ll be. My definition of productivity is greater ability to produce value (not just quantity of outputs). The more value you generate, the greater your impact will be.

(Side note: Is this was the conversations looked like when Google search was introduced? People were raving about how important it was to use Google to search things??)

3 Guidelines for Prompt Engineering

Prompting is only as useful as you make it. Poor prompts result in unhelpful responses. Great prompts result in helpful responses. Simple math.

There are 3 core components I try to include when I want a really high quality response:

  1. Context: The more context, the higher quality response. Who are you? Why are you asking this question? To what end are we exploring this topic? What else about your situation is useful?

  2. Need: The more specific, the higher quality answer. Be specific about what you need the AI agent to do. Create a template for a hard conversation? Break down a complex idea into fundamental pieces? Whatever you need, try to get beyond “tell me about XYZ”.

  3. Tailor the AI response: If you ask questions about a complex topic, it will likely assume you know something about it. Make sure to teach the AI what type of response you are looking for so that it can tailor its response to your needs. Think about: grammar, vocabulary, verboseness, use of analogies, etc.

TPM Use cases (steal and modify my prompts!)

For the most part, these prompts highlight chat-based AI as a starting point. I intentionally avoided giving company-specific details in any prompts because, well, that’s risky. But I did attempt to get specific enough for AI to understand the situation.

I’ve used a variation of each of these in my TPM practice. They have helped me show up strong to my partners across the company.

Accelerate your learning in specific domains

[Context]
I am a technical program manager. My manager has recently asked me to pick up a new strategic initiative in the domain of Identity Access Management. 
I don't have deep experience in this field but will be interacting with many domain experts such as product managers, engineering managers, and IC engineers. We are attempting to scale our current identity access management systems and also add new features.

[Need]
I need to climb the learning curve quickly in this domain so I can be an effective partner.
Your job is to break down identity access management as a domain, and then to help me understand common system design patterns that are used to implement identity access management. I need you to break it down to fundamental components for me to understand.

[Tailor the AI response]
Keep it simple, vocab and grammar assume beginning knowledge of identity but intermediate knowledge of software engineering. Avoid lengthy answers. When appropriate, use analogies to describe things that are more complex. Focus on the fundamental building blocks first, then ask me what I'd like you to do next.

Decision Making Guidance

[Context]
I am a technical program manager. I'm currently driving a large initiative that crosses 8 teams. There is a major decision that needs to be made which involves complex tradeoff considerations. This involves 3 teams. 

[Need]
I need you to help me in a 2 ways:
First, I need you to create a document which can act as a source of truth for all considerations of this decision across all teams. Please include guidance on best practices for documenting complex decision making.
Second, I need you to act as a coach and help me know how to talk to these teams. Initially, just give me some guidelines for how to approach this effectively. There is a lot at-stake and I need you as a sounding board for my approach.

[Tailor the AI response]
Let's start with the first job, decision documentation. Once you've done that, please stop until I say continue with the 2nd job.
Keep the document focused and concise. I don't want to create a novel. Make sure to add headers and have a place where we can document the final decision as well as options and tradeoffs.

Risk Management and Dependency Mapping

[Context]
I am a technical program manager. I've recently picked up a new initiative around coordinating the implementation of GDPR standards. This is important as we scale our companies to international geographies. I've started a partnership with our data architect and a person from compliance, I want to make sure I show up as a value-adding partner.

[Need]
Let's go through a few needs I have:
First of all, I need you to break down GDPR standards for me. I have a baseline understanding but verifying what the fundamentals are would help.
Second, I need you to help me identify common system design dependencies when considering an implementation/adoption of GDPR standards. What parts of the tech stack should be considered? Where is most of the work going to be concentrated? What are some edge cases of the tech stack to consider?
Third, help me identify common risks when implementing GDPR. This should cover technical, legal, and execution risks.

[Tailor the AI response]
I gave you 3 instructions, but only start with the first before moving on. I need you to pause and make sure I'm following before jumping ahead.
I want you to pretend like I'm 12 when describing the legal standards of GDPR. Also, be concise in your answers. Avoid lengthy paragraphs. I'll ask for more detail if needed. Use analogies where appropriate.

Resource

đź“š The TPM Craft Advisor GPT

If you’re not new to using an AI agent (like ChatGPT), then you may also be tired of continually giving context that you’re a technical program manager.

That’s why I created The TPM Craft Advisor. This is my go-to GPT when I’m on-the-job.

To access it, I believe you’ll need an account with OpenAI / ChatGPT plus be a ChatGPT+ member. I’m sorry about that barrier to entry, but just want to prep you for that! If it were me, I would give it to you for no cost but its behind OpenAI paywall!

That’s all folks!

Today’s Sponsor

Do you enjoy learning about the AI space? I’m personally a subscriber of The Rundown AI. Check it out, its free 👇

Stay up-to-date with AI.

AI won’t replace you, but a person using AI might. That’s why 500,000+ professionals read The Rundown– the free newsletter that keeps you updated on the latest AI news, tools, and tutorials in 5 minutes a day.

đź“‹ Quick Poll