It’s been a few months since I started experimenting with AI tools. Coming from a technical background, this shift has been a long-awaited one. I still remember the days when I had to learn PHP and object-oriented programming. Yes, you heard it right... PHP.
As a GPM, I want to share a few practical ways I’m using AI to reclaim my time and sharpen my strategy.
1. My Critical Reviewer
I have an agent setup that reviews my writing and strategy documents. Context is everything here; I provide the agent with a repository of company-wide documentation, presentations, and meeting notes to give it “ground truth.”
The twist: Each agent is prompted to assume the role of a specific senior leader. This allows me to battle-test my narratives against potential executive critiques before I ever hit “send.” For the curious: I’m currently toggling between NotebookLM and Gemini for this.
2. My Meeting Partner
Automated transcripts are a lifesaver, even if the typos (especially on names) are occasionally hilarious. However, transcripts often die in Google Drive. My meeting agent now extracts decisions and action items, assigns owners and dates, and publishes them directly to Confluence. It saves me 30 minutes of administrative “summarization tax” per meeting. Next up: Automating a Slack follow-up via Cursor.
3. My Research Analyst
Whether it’s experimentation results, user research, or customer complaints, the data is only as good as the question you ask. Asking “Identify 3 opportunities” is too vague. Instead, I feed in user segmentation and journey data to get specific. My best use case so far: Cross-referencing feedback against user tenure. I’m leaning heavily on Gemini and Cursor for this deep-dive analysis.
4. The Onboarding Assistant
New joiners are usually buried under a mountain of documentation. I’ve been experimenting with an onboarding assistant that can be pinged for instant answers. NotebookLM has been a standout here, turning dry docs into “podcasts” and structured weekly training plans that feel much more human.
What’s Next?
So far, these experiments have stayed within my immediate workflow and available tools. I’ve also started prototyping with Lovable, but my next frontier is exploring Claude and more complex orchestration to tie these agents together.