Product Growth: This is what a Google AI PM's tool stack looks like — Marily Nika

What the video covers

Published on January 11, 2026, on Aakash Gupta’s Product Growth channel, this conversation features Marily Nika, an AI Product Manager at Google with eleven years of experience and the instructor of a Maven AI PM bootcamp with 150 participants. The format is a podcast-style interview with live tool demonstrations. Nika walks through six tools she uses daily, explaining not just what each does but when and why she reaches for it in a real product workflow.

Who it’s for

Product managers who want to see what an AI-native PM workflow looks like in practice, as opposed to advice about what AI tools exist. The career framing makes it especially relevant for PMs from traditional domains who want to move into AI product roles within twelve to eighteen months.

Key takeaways

1. Prototype before documenting, not after. Nika describes a workflow shift she observes at Google: instead of writing a PRD, sharing it for comments, and aligning stakeholders on paper, effective AI PMs build something interactive first and let stakeholders debate actual functionality. Google AI Studio figures prominently here — she shows building a functional app in under ten minutes as a stakeholder communication tool rather than an engineering output.

2. AI literacy matters more than coding ability. The skills Nika identifies as separating effective AI PMs are understanding data dependency, probabilistic model outputs, APIs, versioning, and what “productionization” means at scale. Writing code is not on the list. The PM’s job remains defining the problem, measuring impact, and translating between user needs and technical constraints — the AI-specific layer is knowing enough about how models behave to set realistic expectations and catch failure modes early.

3. Custom tools compound over time. Nika uses ChatGPT with a Custom GPT trained on her own PRDs to generate first drafts in her writing style. She describes sharing the GPT link with colleagues openly as normalizing AI use rather than hiding it — a cultural point as much as a tactical one. Notebook LM handles research tasks: she describes uploading a four-hour investor relations video and extracting fifteen key points as interview preparation.

4. Move laterally from existing expertise. Career advice for non-AI PMs follows a “be like a crab” framing: move sideways from your current domain into its AI-adjacent product rather than starting from scratch. Examples she gives — a hearing aid specialist becoming an AirPods PM, an ESPN journalist becoming a Meta sports AI PM — illustrate that domain knowledge is an asset, not a liability, when AI enters a field.

5. The AI PM category has a short shelf life. Nika predicts that within two to three years, “AI PM” will dissolve back into standard product management as AI becomes embedded in all product roles. The implication is that the window to build a differentiated AI PM profile is now — not as a permanent specialty but as a transition credential.

Worth watching if

You are a PM looking to audit and update your daily workflow and want to see specific tools in use, not just discussed in the abstract. The live demo component — including a LinkedIn collage generator built in ten minutes using text-to-image models — makes the capability claims concrete in a way that written summaries cannot.