💡

InsightHunt

Hunt the Insights

M

Marily Nika

Product Lead, Metaverse (Avatars & Identity)

Meta

🎯 Product Strategy (1)📈 Growth & Metrics (1)👥 Team & Culture (1)

Key Takeaways

  • 1.Generalist PMs build the right product; AI PMs solve the right problem.
  • 2.Avoid the 'Shiny Object Trap': Do not implement AI just for the sake of using new technology; ensure there is a genuine user pain point.
  • 3.Do not train custom models for an MVP; use prototyping tools or manual operations ('Wizard of Oz') to validate demand first.
  • 4.AI Product Management requires a shift in mindset to embrace uncertainty and non-deterministic outcomes common in research.
  • 5.PMs act as the bridge between Research Scientists and business goals, translating academic hypotheses into monetizable features.
  • 6.Even without a CS degree, PMs should learn the basics of training a model (e.g., using No-Code tools like AutoML) to understand trade-offs and build intuition.
  • 7.Use 'adjacent products' (successful examples from other companies) to get stakeholder buy-in for risky AI bets.

Methodologies(3)

🎯 Product Strategy

A strategic approach that prioritizes the identification of a specific user pain point over the desire to use AI technology. It treats AI as a 'smart solution' that is only deployed after the problem and high-level solution are clearly defined.

Core Principles

  • 1.Step 1: Identify a genuine pain point or problem (don't start with 'Let's use GPT').
  • 2.Step 2: Define a high-level solution (e.g., 'We need to automate X').
  • 3.Step 3: Determine if AI is the 'smart way' to solve it (vs. heuristic rules).
  • +1 more...

"The generalist PM helps their team and their company build and ship the right product. But the AI PM helps their team or company solve the right problem."

#problem-first#discovery#strategy
View Deep Dive →
📈 Growth & Metrics

Instead of building actual ML models for an MVP, PMs should use design prototypes (Figma) or manual 'Wizard of Oz' techniques to simulate the AI experience. This validates user intent and market demand before engineering resources are committed.

Core Principles

  • 1.Principle 1: Zero Code AI - Don't waste data science time on training models for concept testing.
  • 2.Principle 2: Simulation - Use Figma or manual interventions to 'fake' the AI's output.
  • 3.Principle 3: Validate Value - Focus on whether the user *wants* the recommendation/automation, not on how accurate it is yet.
  • +1 more...

"Don't do it for your MVP. It makes zero sense... Just fake what the AI is going to be doing."

#ai-free#validation#growth
View Deep Dive →
👥 Team & Culture

A collaborative framework specifically for AI products that adds a heavy 'Feasibility' bubble driven by research science. The PM acts as the translator between business viability and the non-deterministic nature of scientific research.

Core Principles

  • 1.Desirability (User): Does the user want this personalized/automated experience?
  • 2.Viability (Business): Can we monetize this? (e.g., figuring out pricing for generative AI features).
  • 3.Feasibility (Science): Deep collaboration with Research Scientists. Understanding that models are probabilistic (70% accurate) not deterministic.
  • +1 more...

"You need to get comfortable with having a partner that's a research scientist... working with research, it's more like 'we're going to try this and in a year if it doesn't work out we're going to shut it down'."

#competency#triad#team
View Deep Dive →