Sample AI-native Product Manager Stack, 2026

Harshal
12 hours ago
4 min read

Mapping AI tools to the Product Lifecycle

Nowadays, one of the things I enjoy discussing with product peers is how AI is changing product management. I also have a lot of fun using AI in a structured way, not as random prompts. Here I’ll share how AI is helping me across different parts of the product life cycle, with examples from my day-to-day.

You need 3 minutes to read this.

AI PM workflow including research, analysis, documents, planning, and launch

UX Research

I use AI-enabled user research tools when talking to users or customers. HeyMarvinand Dovetail both have AI capabilities to speed up gathering insights from user interviews and usability studies.

I have used Granola, Fathom-video, and NotyAI to handle meeting notes and insights. They record meetings, create summaries, and organise notes.

Data Analysis

I use AI to write SQL queries based on what I want to analyze. For example, I use Cursor (which has Composer and Anthropic cloud sonnet models), BigQuery extension, and BigQuery MCP to understand what tables are there, what columns are there, and use the following my instructions to write right queries. This overall system also checks for linting. I can even run small queries within Cursor, although I got used to copy-pasting the queries into Metabase or DataStudio. I aim to create a carefully labeled repository of my SQL files, which enables a compounding effect because every new SQL file is written faster by AI by looking at patterns from earlier SQL files.

I also use PostHog, which has a really good AI assistant to understand users' behavior on web pages.

Competitive Research

I use Perplexity Deep Resesarch to do competitive research. I have a template that I use to do competitive research.

Market and AI News

I use n8n to keep track of competitive, market, AI-specific research or news. For example, I have a workflow to extract value from newsletters and blogs and convert them to a private podcast feed to save time and enable me to learn more.

Perplexity and ChatGPT help explain other concepts. I use NotebookLM to understand a topic from many angles, especially complex AI research papers.

Prototyping

I use Lovable, Cursor, and a few more coding agents to prototype products or experiences.

Customer Journey Mapping

I use Mermaid diagrams to create customer journey maps. They aren't the best for a CJM, but Mermaid AI is great, which makes this a fast and hence, preferred approach.

Product Roadmap

I use Notion to keep track of the product roadmap. Its not purpose built for a roadmap, but its DB feature and Notion AI help accelerate its use for roadmapping.

Product Requirements Docs and User Stories

I use Notion with Notion AI to write product requirements documents or JTBDs.

I've also experimented with Cursor and a directory full of organized markdown files so I can easily pass all that context to a much more powerful AI in many more versatile ways than just Notion. This is working great too, but at the cost of readability or sharability.

Task Management

I use n8n, Cursor, Notion MCP, Linear MCP to create and update user stories. I use n8n to automate or maintain a higher quality database in Notion - my n8n workflows read human updated rows to trigger updates to the database.

I have n8n workflows to clean up older tickets or update some status fields in completed tickets.

I use Trello for my side projects or homework, and there I have built an integration using Trello community MCP and built an n8n workflow which I can trigger from anywhere including my phone to create or update tasks. Having an AI assistant is a must-have for project management.

Stakeholder and Customer Communication

I use Loom to tell the user's story to my team, customers, or stakeholders. Loom and Descript both have AI features to help edit videos.

Early customers Onboarding

Along with colleagues, I integrated an AI assistant with Grafana, which helped analyze failures that users faced so that I can provide them a white-glove service. This was most relevant when onboarding early customers.

Missing Tools

I haven't yet used Gamma AI. I know it's really good for presentations. I haven't yet had a need for it. I am happy to be able to use documents instead of presentations most of the time. I haven't used Claude code, but I use coding agents through IDE, web browser, and CLI daily. I have heard Hex is great, but I haven't tried it yet for data analysis. I prefer using vibe-coded solutions for some tasks like Oppotunity-Solution Trees so haven't used Figma and Miro's AI much.