AI, ML, LLM and Generative AI (just rounded to AI on this page for the purposes of brevity) products are technologies that are rapidly sweeping into regular software product development. The following is my musings and discoveries about how the Product Management job changes or stays the same when implementing this technology to solve a problem.

<aside> ⚠️ This is emerging technology, that I’m learning alongside…well.. everyone else. Take these thoughts and opinions in stride.

</aside>

Key Links around the ✨‘book

Managing Data Models

Managing LLM-Based Products

Prompt Hacks

AI-Augmented Product Manager

AI Product-Market Fit

Sources of Inspiration

Managing Data Models details the differences to Product Managing the creation of a data model vs. traditional Product Management

Managing LLM-Based Products details the differences to Product Managing with LLMs vs. traditional Product Management

Prompt Hacks details some pointers that I have learned by experimenting with LLMs

Read the AI-Augmented Product Manager page to understand how I want to use AI to help make my Product Management job more efficient.

The AI Product-Market Fit page is a brainstormed list of the ways that AI can serve as a fit for a customer technology problem.

The Sources of Inspiration is an ongoing list of the media I have consumed that informs how AI affects areas of my career as a Product Manager

Read on below to read my take on the differences in AI Opportunity Discovery and User Experience

AI Opportunity Discovery

10X Thinking

While normally 10X thinking gives me I-don’t-wanna-be-Steve-Jobs vibes, when considering AI solutions, it’s expensive to implement. There’s a ton of governance and infrastructure that goes into a great product, even if someone just gets MS Copilot off the shelf. For the first wave of products, it’s important that the idea and purpose that is being solved can support not only the revenue, but also the innovation budget for the next wave of AI products at the company, especially in the context of enterprises just initially starting on this journey or startups looking to raise funding.

Moats and Secret Sauces

It’s important to analyze just how differentiating AI is making your product and how quickly someone could recreate your product (or a similar product that solves the same pain point).

Helpful questions

• What are the KPIs that demonstrate the success of this product? How do those KPIs tie into revenue for the company?
• How easy can this product be replicated to solve the user need? Is this okay? Why or why not?
  • For example, a document generation product might be easily replicated, but your version is easier because it’s ingrained in their workflow and already trained on their data. So, someone could ask ChatGPT to draft a similar document as a starting point, but it’s harder and won’t be as good (ideally).
• How much margin is expected on this product? Will the solution cover the costs of innovation and infrastructure?
• How much better do you think an ML solution can be? If you think an ML solution might be only a small improvement, that might indicate the current solution is the best one, or a baseline algorithm that’s cheaper to develop, more interpretable (easier to explain to customers) and more simple to maintain than a model.

User Experience with AI