Sign in Subscribe

The Power Shyft: AI for Product Managers

The Power Shyft: AI for Product Managers

Everyone’s using - or at least exploring - AI in their product job. Some have it down to a science now (I am always looking to learn from those folks!). Others are still finding their footing in where and how to best use these tools.

In this article, I want to walk you through some of the everyday ways AI can make you more effective as a product manager. There are certainly many other ways these tools are supersizing our impact, but these are the “low-hanging fruit” that I believe we should all be using by now.

One note before we jump in. I am intentionally not talking about any specific tool. I personally primarily use Claude, Perplexity, and ChatGPT. Each of these tools could help you with all of the work I discuss below. I know there are many others that could as well. The key is to find what works for you. For the purposes of this article, the tool is agnostic to the use cases.

The Five Power Shyfts AI Gives Product Managers

Yeah, I know it’s “shift” but my company’s name is Shyft PM—because things are shifting in product management, and we all need to understand the why (that’s the Y) of that shyft. See what I did there? :)

Anyway, I’ve organized this around what I call the five Power Shyfts—the fundamental changes in how you can work when you have AI as your tool. These aren’t theoretical. I use all of these regularly, and I’ve seen countless product managers do the same.

Power Shyft #1: Infinite Research Assistant

Remember when doing competitive research meant spending hours clicking through websites, taking screenshots, and copying features into a spreadsheet? Or when analyzing customer feedback meant reading through hundreds of support tickets one by one?

AI can do that in minutes.

Here’s what this looks like in practice:

You can feed AI your last 500 customer support tickets and ask it to identify the top themes, segment them by customer type, and highlight any patterns over time. Work that would have taken days happens in seconds.

You can ask AI to analyze your top three competitors’ websites, documentation, and public product updates to create a feature comparison matrix. It won’t catch everything (you still need to actually use the products), but it’ll get you 80% of the way there as a starting point.

You can upload the recording or transcript of a customer interview and have AI pull out the key pain points, identify contradictions, flag areas that need follow-up, and even suggest questions for your next interview.

The key here is that AI doesn’t replace customer research. It accelerates and scales it. You still need to talk to customers. You still need to observe them using your product. But AI can help you find patterns across conversations, analyze data from multiple sources, and surface insights you might have missed.

Power Shyft #2: Instant First Draft Generator

Here’s a truth about product management: we spend a shocking amount of time writing things.

Product briefs. One-pagers. Roadmap presentations. User stories. Feature specs. Release notes. Stakeholder updates. Competitive analyses. PRDs. Strategy docs. Email updates. Slack messages explaining why we’re not building that feature right now.

So. Much. Writing.

And here’s another truth: the blank page is the enemy. Starting from scratch is hard. Staring at a blinking cursor wondering how to structure a product brief is not the highest and best use of your cognitive energy.

AI solves this.

Give AI your basic requirements, and it’ll create a first draft. Not a final draft. Not something you’d ship as-is. But a structured starting point that you can edit, refine, and make your own.

Here’s what I mean:

For product briefs: Tell AI the problem you’re trying to solve, who it affects, what success looks like, and any constraints you’re working with. It’ll give you a structured brief with problem statement, user needs, success metrics, scope, and risks. You’ll need to edit it—add your insights, adjust the framing, inject your strategy. But you’ve saved yourself an hour of staring at a template.

For user stories: Describe the feature and the user outcome. AI will draft the stories in whatever format you use (Gherkin, job stories, traditional user stories). Again, you’ll refine them, but you’ve got the structure.

For release notes: Give AI the list of features and bug fixes. It’ll turn them into customer-friendly language. You’ll need to adjust the tone and add context, but the mechanical work is done.

For status updates: Tell AI what you shipped, what’s in progress, what’s blocked, and what decisions you need. It’ll structure it for whatever audience you need—executives, the team, customers.

If you spend 2-3 hours every week writing stakeholder updates, you can reduce that to 20 minutes feeding AI the key points and then editing its output to emphasize what matters. That’s 2+ hours back in your week to do other product work.

Important caveat: AI-generated content is sometimes obvious. Each one of us has our own writing “voice”. So don’t just copy-paste. Use it as a starting point, then make it yours. Add the personality. Add the context. Add the judgment.

Power Shyft #3: On-Demand Thought Partner

This one’s subtler but might be the most valuable.

Product management can be lonely. You’re often the only PM on a product, or you’re working on a gnarly problem that’s uniquely yours. You need to think things through, but you don’t always have someone to think them through with.

AI can be that thinking partner.

Here’s how I use this:

When I’m stuck on a problem: I’ll describe the problem to AI and ask it to suggest different ways to frame it, different approaches I might not have considered, or what questions I should be asking that I’m not asking yet. It’s like rubber-duck debugging, except the duck talks back and sometimes says useful things.

When I’m preparing for a difficult conversation: I’ll describe the situation and ask AI to role-play the other person. “I’m about to tell Engineering that we need to deprioritize their favorite feature. You’re the Engineering Lead who’s been pushing for this feature for six months. Push back on my reasoning.” This helps me pressure-test my thinking and prepare for objections.

When I need to explore options: I’ll ask AI to help me think through different approaches to solving a problem. “What are three different ways we could address this customer need? What are the pros and cons of each?” It won’t give me the answer, but it’ll help me think more broadly.

When I’m analyzing data: I’ll share metrics or user research findings and ask AI to help me identify patterns, suggest what might explain what I’m seeing, or propose what additional data would be useful to gather.

Think about the last time you got stuck choosing between two high-impact features. Instead of going in circles solo, you could spend 30 minutes with AI talking through each option—the tradeoffs, the strategic implications, all of it. AI asks questions you hadn’t considered. Identifies assumptions you’re making. Maybe even suggests a third option that gets you unstuck.

Will AI make your decision? No. That’s still your responsibility. But it can help you think more clearly about the decision. And sometimes that’s exactly what you need.

Power Shyft #4: Audience Context Translator

You know what’s exhausting? Explaining the same thing fifteen different ways to fifteen different audiences.

The engineers need the technical details. The executives need the business case. Sales needs the customer pitch. Marketing needs the positioning. Customer success needs the implementation guide. And they all need it explained differently, at different levels of detail, with different emphasis. This has always been one of the toughest things that product managers must do… and it’s what can separate the great from the good.

Well, now AI can help us all with this translation work.

Give AI your core message and tell it who the audience is. It’ll reframe it appropriately.

“Take this technical explanation and turn it into something I can share with executives in 3 bullets.”

“Here’s our product strategy. Help me turn this into talking points for the sales team that focus on customer outcomes, not features.”

“This is the feature we’re shipping. Create an FAQ for customer success that addresses the top 10 questions they’re likely to get.”

I was working with a team that had built a sophisticated machine learning feature for fraud detection. The PM understood how it worked, but she was struggling to explain it to different audiences.

We used AI to create five different versions:

  • A technical deep-dive for the engineering blog
  • A business case for executives focused on ROI and risk reduction
  • A customer-facing explanation that emphasized safety without scaring anyone
  • A sales enablement guide with objection handling
  • A simple FAQ for support

Each version was different. Each emphasized different aspects. And the PM edited each one to ensure accuracy and add context. But instead of writing five documents from scratch, she had five starting points. Time saved: probably 6-8 hours.

The goal isn’t to outsource your voice or your context. It is to help you communicate what you already know in a way that will better resonate with each audience.

Power Shyft #5: Pattern Spotter Across Everything You’re Doing

This last one is less about specific tasks and more about meta-awareness.

AI can help you spot patterns across your work that you might not notice because you’re too close to it.

Feed AI your last 10 product decisions and ask it if there are any biases or patterns in how you’re prioritizing. You might discover you’re consistently overweighting one customer segment, or that you’re avoiding certain types of technical debt, or that you’re not investing enough in platform improvements.

Share your last month of customer conversations and ask AI what themes are emerging across all of them. Sometimes the pattern isn’t obvious when you’re in the individual conversations, but it becomes clear when you look at them all together.

Give AI your roadmap from the last year and ask it to identify what categories of work you’re doing most and least. You might realize you’re spending 80% of your resources on incremental improvements when your strategy says you should be making bigger bets.

This isn’t about outsourcing your judgment. It’s about getting a different perspective on patterns in your own work. Think of it like having someone audit your work and point out tendencies you might not see yourself.

The Daily Practice: How This Actually Works

So what does this look like day-to-day?

Here’s my typical week now, with AI integrated:

Monday morning: (note: I wish I could honestly say I do this every week, but I get lazy more often than not. But when I do it, it’s a great help.)

I review my calendar and think of my goals for the week and then use AI to help me draft my prioritized to-do list. I am more organized and effective throughout the week when I do this.

During customer interviews: I record the conversation (with permission), then feed the transcript to AI afterward to pull out key quotes, identify pain points, and suggest follow-up questions. This saves me an hour of note-processing per interview.

Before important meetings: I brief AI on the context and ask it to help me think through potential objections, questions I should be ready for, and how to structure my argument. For the most important meetings, I also practice by pretending my cats are my audience. They love that. I know that’s not AI related but it helps, and I highly recommend it.

When I’m stuck: I talk through the problem with AI. Sometimes it helps me get unstuck. Sometimes it just helps me articulate what I’m thinking more clearly.

Throughout the week: Anytime I’m writing something—a brief, a spec, an email, a Slack message—I consider whether AI can give me a head start. Usually it can.

Am I using AI every minute of every day? No. Do I trust it blindly? Definitely not. But do I use it multiple times a day to work faster and think better? Absolutely.

What AI Doesn’t Do (And Why That Matters)

Let me be crystal clear about what AI doesn’t do for product managers.

It doesn’t:

  • Make your decisions for you
  • Understand your customers better than you do
  • Build relationships on your behalf
  • Know your company’s strategy and constraints
  • Exercise judgment in ambiguous situations
  • Replace the need for you to think deeply about problems

AI is an accelerator, not a replacement. It makes you faster at gathering information, drafting content, and processing data. But it doesn’t change the fundamental work of product management, which is synthesis, judgment, and human connection.

The best product managers I know use AI heavily. But they’re also deeply engaged with customers, thoughtful about strategy, and careful about decisions. They’ve just freed up time from the mechanical work to focus on the strategic work.

That’s the goal. Not to let AI do your job, but to let AI handle the parts of your job that don’t require your unique human judgment, so you can focus on the parts that do.