How Product Managers Are Using AI to Automate the Tasks They Hate
What You'll Find In This Article
- •Understand why personalized AI copilots outperform generic ChatGPT queries for real work tasks
- •Identify which repetitive tasks in your workflow are good candidates for AI automation
- •Learn how to give AI tools the context they need to actually understand your business
- •Recognize the difference between using AI as a search engine versus building AI workflows that compound knowledge
If you've tried ChatGPT and thought "that's cool, but how does this actually help my day job?"—you're not alone. Most people hit a wall because generic AI queries don't know anything about your company, your customers, or what you're actually working on.
The game-changer, according to a new piece from Lenny's Newsletter, is building AI assistants that remember your context over time. Instead of starting from scratch every conversation, these "copilots" learn your company's quirks, track your ongoing projects, and handle the repetitive work that eats up your calendar—like summarizing customer feedback, monitoring competitor pricing, or turning feature requests into organized reports.
The article argues this isn't about replacing product managers. It's about freeing them to focus on the uniquely human stuff: understanding customers, building relationships, and making judgment calls. The teams pulling ahead aren't just using AI—they're building AI workflows that get smarter over time.
The Problem
Product managers are drowning in busywork. Every day brings a flood of customer feedback to read, competitor moves to track, feature requests to organize, and stakeholder updates to compile. These tasks are important but mind-numbing—and they crowd out the strategic thinking that actually moves products forward.
Meanwhile, most people's experience with AI tools like ChatGPT has been... underwhelming for actual work. You can ask it trivia questions or help draft an email, but it doesn't know your company, your customers, or what you're working on. Every conversation starts from zero.
The Solution Explained
The breakthrough is moving from "AI as a search engine" to "AI as a personalized assistant that learns your business."
Think of it like the difference between asking a random stranger for directions versus having an assistant who's been working with you for months. The stranger can give generic advice. Your assistant knows your schedule, your preferences, your ongoing projects—and can actually take action on your behalf.
New tools now let you build these personalized AI copilots without writing code. You can give them access to specific documents, teach them your company's terminology, and have them remember context across conversations. Over time, they get better at understanding what you need.
How It Actually Works
Step 1: Give your AI context about your work Instead of generic prompts, you connect your AI assistant to relevant documents—your product roadmap, customer personas, company style guides. This becomes its "knowledge base" about your specific situation.
Step 2: Set up automated monitoring Using tools like Slack integrations or dedicated platforms like Lindy AI, you can create "agents" that watch for specific things: new customer complaints in your feedback tool, price changes on competitor websites, mentions of your product on social media.
Step 3: Let AI handle the synthesis When information comes in, your AI doesn't just alert you—it summarizes what matters. Instead of reading 50 customer reviews, you get a daily digest highlighting the three recurring themes and any urgent issues.
Step 4: Build compounding knowledge The real magic happens over time. Your AI remembers past conversations and builds on them. Ask about customer feedback in January, then follow up in March asking "how has this changed?"—and it actually knows the context.
Real Examples
Customer feedback on autopilot: One team set up an AI agent that monitors NPS survey responses. When a customer leaves negative feedback, it automatically creates a support ticket in Zendesk, categorizes the issue, and flags it for the right team member.
Competitor intelligence without the spreadsheets: Instead of manually checking competitor websites, an AI agent monitors pricing pages and sends a Slack alert whenever something changes—complete with a summary of what's different.
Rapid prototyping: AI design tools can now generate brand-aligned mockups in under five minutes. A product manager can describe a feature idea and see a clickable prototype before their next meeting—no designer required for the initial exploration.
Feature request organization: Rather than manually tagging and sorting incoming feature requests, an AI copilot reads each one, categorizes it by theme, estimates complexity, and adds it to the appropriate spot in your tracking system.
Choose one repetitive task you do weekly (like reading customer feedback or checking competitor sites)
Pick a tool: ChatGPT Plus for basic copilot features, or Lindy AI for more advanced automation
Gather 3-5 documents that explain your context (product overview, customer personas, competitor list)
Upload documents to your AI tool and write a short prompt explaining what you want it to help with
Test with a real task and refine your instructions based on what works
Set up one automated monitoring task (competitor alert or feedback digest)
PROMPT:
"What's one task I do every week that's important but tedious—and could be handled by an AI that knows my company?"