How to Be a Product Manager AI Can't Replace
Why the PMs who survive the AI wave won’t be the ones who ship fastest. They’ll be the ones with taste, judgment, and the spine to use them.
👋 Hey, Sam here! Welcome back to The Product Trench. I started this newsletter to cut through the noise and share actionable insights, no-nonsense advice, and stories related to product management and leadership. Occasionally, I share hot takes on topics that get me fired up.
Was this forwarded to you? Join The Product Trench 👇
It’s been a while! No big reason, just life doing its thing. Work got intense, we added a tiny human to the family (sleep is a distant memory), and somewhere in there, I realized I needed to actually take care of myself for a change. Physical health, mental health, the whole deal.
What I did miss, though, was having a place to work through ideas that have been rattling around in my head. And lately, one idea keeps coming back: the difference between PMs who just execute and PMs who actually give a shit about what they’re building.
So here we are.
My therapist once told me I have “professional FOMO.” He meant it as a diagnosis, but honestly, it felt more like a badge of honour. I spend an unhealthy amount of time thinking about why some PMs light up a room when they talk about their product, while others seem to be running out the clock until their next sprint planning meeting.
A few years back, I watched a PM present a feature launch in an all-hands. Great slides. Clean metrics. Everything shipped on time. But when someone asked, “Why did we build this?” there was this awkward pause before she said, “Well, leadership asked for it.” The room got quiet in that specific way that makes everyone suddenly very interested in their laptops.
That moment stuck with me because I’ve been that PM. Early in my career, I treated product management like project management with better snacks. Build what’s asked, ship it, measure it (maybe), then move to the next thing. I was a feature factory employee with a fancy title. And the thing is, nobody explicitly told me I was doing it wrong. I hit my OKRs. Stakeholders seemed happy. But I was basically a well-organized vending machine.
The Opinions Deficit
Here’s what nobody tells you in those “So You Want to Be a PM” blog posts: the job is fundamentally about having a point of view and defending it with evidence. Not in an egotistical “I know best” way, but in a “I’ve thought deeply about this and here’s what I believe” way.
The PMs who treat the role like advanced project coordination are everywhere. They take requirements from sales, engineering, leadership (whoever shouts loudest) and turn them into Jira tickets. They run standups like traffic cops. They launch features and immediately forget they exist because there’s another deadline breathing down their neck.
I’m not being harsh for sport. I genuinely believe this happens because companies accidentally train PMs to be order-takers. When your performance review focuses on “delivery velocity” and “stakeholder satisfaction,” you learn fast that having opinions creates friction. Friction slows things down. Slower is bad. So you become the path of least resistance.
But here’s the problem: when AI can write your PRD, analyze your user data, and probably run a decent sprint retrospective, what’s left that makes you valuable?
Agency. That’s what’s left.
What Agency Actually Looks Like
Agency isn’t about being the loudest person in the room or having an opinion on everything. I’ve worked with PMs who confuse contrarianism with critical thinking. They’ll argue about button colours for 20 minutes because they think that’s what “being opinionated” means.
Real agency shows up differently. It’s the PM who ships a feature and then obsessively checks the data for the next two weeks, not because they’re measured on it, but because they genuinely want to know if they were right. It’s the person who speaks up in a leadership meeting when everyone’s nodding along to a bad idea because they’ve actually talked to customers and know it won’t work.
I once worked with a PM (let’s call her Sarah) who managed our accounting integrations. During a quarterly planning session, leadership wanted to prioritize expanding our QuickBooks integration to support inventory tracking, job costing, and project management features. Sounded reasonable. Our enterprise customers had been asking for it. Everyone was ready to commit the quarter to it. Sarah raised her hand and said, “I don’t think we should build this.” You could feel the temperature drop.
She pulled up her laptop and showed us data from the past six months. Eighty-seven percent of our QuickBooks users only synced invoices and expenses. The remaining thirteen percent who used advanced features were mostly on custom enterprise plans with dedicated support. She’d also talked to eight customers the previous week who said they wanted “better QuickBooks integration.” When she dug deeper, they meant the basic sync was breaking, not that they needed more features.
We didn’t expand the integration. Instead, we spent the quarter making invoice and expense syncing bulletproof. Boring. Unsexy. But support tickets dropped significantly, and we stopped losing customers over sync failures.
That’s agency. Not being contrarian for its own sake, but doing the work to form an opinion and having the guts to share it when it matters.
The Curiosity Crisis
The PMs who struggle with agency often have a curiosity problem. They don’t seem genuinely interested in whether what they built worked. Launch happens, maybe there’s a Slack celebration with some emoji, and then it’s on to the next thing.
This is wild to me because the most interesting part of product management happens after launch. That’s when you find out if you were right. That’s when customers do weird, unexpected things with your feature. That’s when you learn something that changes how you think about the problem.
But if you’re not curious, you miss all of it. You treat launches like closing tickets instead of opening questions.
I’ve started asking PMs in interviews: “Tell me about something you built that failed.” The ones without agency give you a well-rehearsed story about a failure that wasn’t really their fault (bad timing, lack of resources, stakeholder interference). The ones with agency get animated. They’ll talk about what they learned, how it changed their thinking, and what they’d do differently. They’re still processing it, even if it happened years ago.
Curiosity isn’t a personality trait you either have or don’t. It’s a muscle. You build it by asking one more question in customer interviews. By checking the dashboard on Saturday morning, even though nobody’s making you. By reading your competitors’ release notes, you want to understand their thinking.
The Meeting Room Test
Here’s a simple diagnostic: pay attention to who talks in meetings. I don’t mean who talks the most (we all know that person, and they’re usually not adding much value). I mean who contributes signal.
PMs with agency ask clarifying questions. They push back on assumptions. They connect dots between things that seemed unrelated. They’re not performing; they’re actually engaged in figuring something out.
PMs without agency are silent until someone asks them directly about the timeline or scope. Then they give a careful, noncommittal answer that keeps everyone happy. They’re optimizing for not making waves.
The frustrating part is that silent PMs often think they’re being professional. They’re “staying in their lane” or “letting others drive the vision.” But product management isn’t about staying in your lane. The job is to have an informed opinion about whether we’re going in the right direction and to say something if we’re not.
I remember early in my career, staying quiet in executive meetings because I figured the VPs knew better than me. Then one day, my boss pulled me aside and said, “You’re closer to the customer than anyone in that room. If you’re not speaking up, we’re making decisions with incomplete information.”
That reframed everything. My opinions weren’t me being full of myself. They were data points the team needed.
Evangelism as a Core Competency
The best PMs I’ve worked with are shameless evangelists for their products. Not in a used-car-salesman way, but with genuine enthusiasm about the problem they’re solving and the people they’re solving it for.
They bring up their product in random conversations. They forward customer feedback to the team even when nobody asked for it. They write detailed launch announcements that people actually read. They make their engineers feel like they’re working on something that matters.
This isn’t about being extroverted or charismatic. I’ve known quiet PMs who were incredible evangelists through written updates and one-on-one conversations. It’s about caring enough to make other people care.
When you don’t evangelize, you’re sending a message: this work isn’t worth getting excited about. Your team picks up on that. So do stakeholders. So do customers.
The PM I mentioned earlier, who couldn’t answer “why did we build this”? I don’t think she was lazy or incompetent. I think somewhere along the way, she learned that caring too much made the job harder. Easier to just execute and move on. But that’s a terrible way to spend your career, and it produces mediocre products.
The Cultural Dimension
Some of this is on the PM to fix. But let’s be honest: company culture plays a huge role in whether PMs develop agency or learn to be quiet.
If your performance reviews focus exclusively on output metrics (features shipped, velocity points, roadmap adherence), you’re training PMs to be delivery managers. If “stakeholder satisfaction” is a key metric, you’re training them to be people pleasers. If you celebrate launches but never talk about outcomes, you’re teaching them that what happens after doesn’t matter.
I’ve seen companies accidentally create cultures where having opinions is risky. Where challenging a VP’s pet project is career-limiting. Where data that contradicts the roadmap gets ignored. In those environments, the rational move for a PM is to keep your head down and ship what you’re told.
But those companies are sleepwalking into irrelevance. Because as AI tools get better at the tactical parts of product management, the only differentiation left is judgment, taste, and the willingness to say hard things.
If you’re a PM in one of these cultures, you have a choice. You can adapt to it and become really good at going through motions. Or you can start small (speaking up in one meeting, digging into one launch’s data, asking one tough question) and see what happens.
Sometimes the culture is more flexible than it seems. People are waiting for someone to model a different way. Sometimes you’ll get shot down and learn that this company isn’t the place for you. Either outcome is better than spending years being quiet.
The AI Accelerant
We’re at an inflection point. Tools like ChatGPT, Claude, and ChatPRD (they’re already decent at writing PRDs, analyzing data, and generating ideas). They’ll get better fast. The parts of product management that feel like sophisticated busywork are automatable now.
What’s not automatable is the PM who understands the second-order effects of a pricing change because they’ve spent hours talking to customers. The one who knows which engineer on the team is burning out and needs support. The one who can read between the lines of what a stakeholder is saying and address their actual concern.
These skills all require agency. You can’t outsource having a point of view.
I’ve started using AI tools for a lot of tactical work. They’re great at first drafts, data analysis, and brainstorming. But they can’t tell me if an idea is worth pursuing. They don’t have intuition about what will resonate with our specific users. They don’t wake up at 2 am thinking about a problem.
The PMs who lean into AI as a tool to free up time for higher-order thinking are going to run laps around the ones who see it as a threat. But you have to actually use that freed-up time to think, to develop opinions, to engage deeply with the problem. If you’re just using it to ship more features faster, you’re missing the point.
Strong Opinions, Loosely Held (For Real This Time)
That phrase has become a cliché, but I still think it’s the right frame. The “strong opinions” part means you’ve done the work (talked to users, looked at the data, thought about the market) and arrived at a point of view you can defend. The “loosely held” part means you’re willing to change your mind when you encounter better information.
What doesn’t work is having no opinions because you’re waiting for perfect information, or holding your opinions so tightly that you ignore contradictory evidence.
I’ve been wrong about products plenty of times. I pushed for a feature that customers didn’t care about. I dismissed an idea that turned out to be great. I misread market timing on a launch. Each time, having strong opinions made it clear I was wrong. And being willing to update those opinions meant I could adjust course.
The PMs who struggle aren’t the ones who are wrong sometimes. They’re the ones who never commit to a position in the first place, so they can’t learn from being wrong.
What This Means for You
If you’re reading this and thinking, “I might be that quiet PM,” I get it. Developing agency feels risky. It’s easier to stay in the background, execute what you’re told, and collect your paycheck.
But I promise you’ll be more satisfied, more valuable, and more resilient in your career if you start building that muscle now.
Start small. Pick one thing you’re working on and develop a real point of view about it. Not just “this seems fine” but “here’s why I think this will work, and here’s what I’d expect to see in the data if I’m right.” Share that opinion with your team. Then check the data after launch and see if you were right.
Ask one more question in customer interviews. Push back once when a stakeholder asks for something that doesn’t make sense. Write a post-mortem that actually analyzes what happened instead of just listing what shipped.
These small acts of agency compound. You’ll get more comfortable having opinions. Your team will start seeing you as someone with valuable judgment. You’ll find the work more interesting because you’re actually trying to figure things out instead of just executing tasks.
And if your company culture genuinely punishes this kind of thinking, you’ll learn that too, which is painful, but valuable information about where to spend your career.
The product management field is professionalizing fast. The bar is rising. The PMs who make it won’t be the ones who can run a sprint or write a PRD. Those will be table stakes that AI handles. It’ll be the ones who have taste, judgment, and the spine to use them.
So yeah, be opinionated. Your products will be better for it. Your team will appreciate it. And you might actually enjoy this weird job we’ve chosen.
👋 And that’s a wrap folks. Thank you for reading.
If you enjoyed reading today’s newsletter, feel free to forward it to someone! Or you can always hit the like button on this post so more people can discover it on Substack.
See you next time.
— Sam ✌️