Today’s episode

Most PMs are using AI the same way they used Google in 2005.

Type something in. Get something out. Move on.

That is not how the best PMs are using it. The best PMs have stopped treating AI as a search engine and started treating it as a team member. One that already knows their product, their writing style, their strategy. One that does not need to be briefed from scratch every single time.

That shift is what today’s episode is about.

I sat down with Lisa Huang, SVP of Product at Xero, an $18 billion finance platform. She built the AI assistant for the first generation Meta RayBan smart glasses. She created Gemini Gems at Google. She has been an AI PM at Apple, Meta, and Google - three of the most demanding AI product environments in the world.

She gave us a masterclass across Gemini Gems, building AI into hardware, running AI agents at scale inside a financial product, and what the AI PM career actually looks like from here.

In today’s episode, we discuss across three topics.

* How to build Gemini Gems and AI projects that actually work.

* What she learned building AI into a wearable device.

* What the future of the AI PM career actually looks like.

----

Check out the conversation on Apple, Spotify and YouTube.

Brought to you by - Reforge:

Get 1 month free of Reforge Build (the AI prototyping tool built for PMs) with code BUILD

----

If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, DeepSky, Relay.app, Magic Patterns, Speechify, and Mobbin - grab Aakash’s bundle.

----

Key takeaways:

1. Stop briefing your LLM from scratch every time - Gemini Gems hold your context permanently. Your role, your company strategy, your writing style. Build it once and it already knows everything the next time you open it.

2. Every PM needs 3 Gems - A writing clone trained on your PRDs and emails. A product strategy advisor loaded with your company docs and competitor analysis. A user research synthesizer that ingests raw transcripts and surfaces key themes.

3. Vague instructions are the number one mistake - "Help me write better" gets you nothing. Write a full page of context. Your role, your audience, your format preferences. The more specific, the more personalized the output.

4. Gemini Gems vs ChatGPT custom GPTs - OpenAI framed GPTs as an app store ecosystem. Google focused on personal productivity instead. First principles beat copying a competitor's framing, and the GPT store never took off.

5. On-device AI is the future for wearables - Cloud is the default today but once a device is on your face all day, people want their data staying local. Privacy beats performance when the device is that personal.

6. Accuracy is the product in high-stakes AI - LLMs out of the box are not great at math, accounting, or tax. Winning agents combine deep domain knowledge with proprietary data that no general-purpose model can access.

7. Measure agents in three layers - Quality first (evals, human annotators, LLM judges). Product metrics second (adoption, retention, CSAT). Business impact third (revenue attribution, ARR). Skip to layer three without the foundation and you are measuring on sand.

8. AI will not replace PMs - it will replace the execution work. Writing PRDs, creating mocks, managing roadmaps. What stays is product judgment. The ability to look at ambiguous signals and say this is the right bet and here is why.

9. The PM role is becoming a hybrid - PM to engineer ratios will compress. The expectation is that PMs also build. Not just spec and hand off, but prototype, design, and code enough to show what they mean. The tools to do this exist right now.

10. Your company's permission is not required - Most companies are not fine-tuning models. They are using the same consumer tools you already have. Build Gems. Build projects. Build small AI products with your personal data. There is nothing stopping you.