Pivoting into AI Product Management: Lessons from Amazon PMs Bhoomika Ghosh and Shrinedhi Rajan
Share
Artificial Intelligence is transforming how we build, test, and scale digital products. For product managers, the real challenge isn’t deciding whether to adopt AI—but understanding how to pivot into this rapidly evolving space.
That’s exactly what we unpacked in our recent virtual fireside chat with Bhoomika Ghosh and Shrinedhi Rajan, both Senior Product Managers at Amazon. Hosted by Products by Women, the session brought together a curated group of product managers, engineers, and data professionals from companies like Google, NVIDIA, Microsoft, PayPal, Oracle, FedEx, and more.
Over the course of 60 minutes, we explored what it means to become an AI Product Manager today, with real-world advice and behind-the-scenes stories from Amazon’s own teams.
What Is AI Product Management?
Shrinedhi Rajan broke down the different ways PMs can work with AI using a framework by Marili Nika:
- AI-Enhanced PMs: Use AI tools to boost their own efficiency.
- AI Experience PMs: Use AI to improve the customer experience (e.g., personalization).
- AI Builder PMs: Collaborate with scientists to build and tune foundational models.
“You don’t need to be deeply technical to work in AI,” shared Bhoomika. “What matters most is your ability to lead with product thinking.”
Build Before You're Hired
Bhoomika showed us just how accessible AI prototyping can be.
She built a fully functional AI-powered gamified learning app—in just a week—using Gemini, Vizly AI, and Replit. The app helps close the gender gap in STEM by allowing users to upload learning material and automatically generate quizzes and content.
“Even small side projects can help you break into an AI PM role. Start where you are,” she advised.
Core PM Skills Still Matter
Both speakers emphasized: You don’t need to be a machine learning expert to be effective. Instead, focus on:
- Framing ambiguous problems
- Leading cross-functional teams
- Thinking in terms of customer impact
- Translating between science and business
AI in Action: Amazon’s Search Autocomplete
Shrinedhi walked us through how her team at Amazon applies AI to optimize search autocomplete.
She explained the difference between:
-
Traditional ML: Uses past data to rank and recommend
-
Generative AI: Synthesizes new responses using foundational models
“Our job is to frame problems as ML opportunities—and define clear success metrics and guardrails.”
Tools and Tips
Not sure where to start? Try this:
- Use Gemini or Claude to analyze customer feedback or draft docs
- Build a side project using tools like Replit, Vizly AI, or Glide
- Follow science blogs by Amazon, Google, Meta for real-world insights
- Learn how to evaluate model performance (accuracy, latency, cost)
Takeaways for Aspiring AI PMs
- Start small: Internal side projects or solo builds show initiative
- Be customer-first: Every AI feature should solve a real problem
- Get comfortable with ambiguity: Working with scientists means co-creating specs, not just handing off tickets
- Understand model impact: You don’t need to build models—but you do need to understand how they affect the user experience
Final Thought
“You don’t need to chase AI for the buzz. Start by asking: what’s the customer problem, and how can AI meaningfully solve it?” — Bhoomika Ghosh, Sr. PM, Amazon
Want to Speak or Sponsor Our Next Event?
Whether virtual or in-person, we’re always looking for thoughtful voices and aligned brand partners. Fill out this form to apply.
For press, partnerships, or sponsorship inquiries, email us at info@productsbywomen.com.
Let’s build the future of product and tech, together.