How Do I Move From Execution to Strategic Influence in the AI Era?

Watch the Episode

Transcript

Ready to move to the next level in your product career? I'm Shobhit from Intentional Product Manager. Join me as we discuss ways to help you stand out in your job search and your career so you can have more impact and make more money. Hey everyone, welcome back to the Intentional Product Manager podcast. I'm your host Shobhit Chugh, and today we have someone whose resume reads like a playbook for world-class product leadership. Mick is a friend from business school — he's worked across Amazon, Microsoft, Meta, Airbnb, SoFi, and now is Head of Product for GenAI Monetization at Scale AI. Mick, welcome to the show, it's great to have you here.

Mick: Thank you so much for having me. It's been too long since business school — and between me and your Google experience, we do have a lot of companies covered.

Shobhit: Exactly. Mick is not just a leader in cutting-edge AI products — he's also built startups, led growth at fintech, brought structure to chaos at some of the biggest names in tech. Let's start with a warm-up question. You've worked at big tech, startups, consulting — what's one experience that fundamentally shaped how you think as a product leader?

Mick: There's two things I've learned out of this experience. The first is when I was a senior IC product manager at Facebook, now Meta, we ran a quarterly product review process — at the beginning of a quarter you present the plan to the VP, but to be ready for that meeting you meet with each of their direct reports individually, address their concerns, nominate one to pre-wire the actual execs, make sure the exec is on board, and then you can actually have the meeting. It's efficient, these execs' time is valuable, but I enjoy building and ideating a bit more than what Strauss would call "product theater" — choreographing that performance. Jeff Bezos talks about how the best ideas have a clean doc with a clean recommendation, but he's also good with messy meetings where people actually talk through ambiguous problems with many potential levers. That Facebook experience made me pivot to smaller companies like SoFi, Brex, and Scale, where I still do stakeholder management, but on balance I'm spending a lot more time building. The second thing is how powerful it is to look back at these companies and see their culture, the pros and cons. Netflix is very metrics-oriented — great quarter, you get promoted, bad quarter, there are implications. You don't need great leaders to say "you exceeded your revenue threshold," you need great leaders to say "in this macro environment, that was actually good." Meta is very relationship- and metrics-oriented — it's important for the heads of eng, PM, and design to get along and calibrate each other's metrics. Google feels much more about being a self-starter and having research breakthroughs — great at long-term tech innovation, maybe less so on the product front, though they've gotten a lot better. I appreciated the long-term thinking at Google more than anything — it was almost not interesting for anyone to know what I was working on over the next year until there was alignment on what it could mean five years out. Are we going somewhere meaningfully different?

Shobhit: Setting vision should be one of the most interesting parts of being a product manager — no execution, no feasibility, just what's the beautiful nirvana you want to get to.

Mick: Exactly.

Shobhit: You've mentioned all these different places — what's something people get wrong about your journey when they look at your resume?

Mick: I think when people look at my resume they see a lot of momentum — joining new companies, learning new products, being a generalist, all exciting and fun. But something I don't think I appreciated enough, even after business school, is how hard it is to embed yourself in a new organization — and this gets much harder the more senior you get. When you join a large company at a senior role, there's a lot of other people competing for the same sets of power, explicit or not, and you need to get through the first six months successfully. Say you're a great interviewer, you join a new company tomorrow based on your own skills — day one you know zero about how the company actually operates and all the research they've done, and you have no network to fall back on. It's really important to have a good manager to help set you up for success in those first six months, and after that you have the network and knowledge to manage your career yourself. My advice: diligence your manager and leverage them effectively. Create a group of peers who can review your ideas and give you feedback, especially if they've been at the company a while. Map out incentives, figure out who the decision-makers are, and definitely try to ship something within the first month — both as a sign you're driving outcomes, and because you learn so much doing the whole end-to-end development process. Creating requirements is very different from doing visual QA, launching things, finding bugs — the sooner you get through that entire cycle, the sooner you know what game you're playing.

Shobhit: I love that last lesson — get through the entire life cycle as soon as possible, you discover things quickly versus holding back.

Mick: Exactly, and to do that early you have to pick something very small — you're not going to ship a fifty-person product that way. Be intentional about that goal: find something really small, make it end to end.

Shobhit: Now let's dig into topics relevant to your career. There's a lot of discussion right now about AI — I recently did an end-of-year podcast recap and my platform automatically picked out the most-used word, and it's literally people saying "AI, AI, AI" over and over. When you look at where AI product management is headed, what do product managers need to unlearn?

Mick: Two things to set context. First, we are so lucky to be in tech and alive at this time — you can read sci-fi from hundreds of years ago about artificial intelligence, and it's actually happening during our lifetimes. If you're able to work adjacent to AI, you should have a lot of gratitude for that. Second, I don't think AI specifically changes the job that much. It's a very valuable tool that can make things that took a long time before much easier and quicker to do. For example, managing the pay team at Scale AI, which pays out contributors across the globe — I can look at all the tickets that came in last year, about 100,000 of them, feed that into an LLM, and it'll tell me the three most common problems with no effort. That would have been a bunch of time in Excel wrangling before. The PM job hasn't changed much — you just have better tools. It's still on the PM to do two things: pick the right problem — one that solves a major user concern, that will make money, that's aligned with your core competencies and where you can differentiate — and align the organization and rally the troops behind a vision, which is very hard for AI to do, that social touch. What is changing is that as a PM you can't do everything yourself, and now you have two ways to solve that: delegate to people you trust, which was always available, and now also lean on an AI tool that speeds things up — though it requires continuous investment making sure the AI has access to all the right context about your company. What PMs need to unlearn is "there's this problem, I'm going to go solve it myself" — now you either delegate or give it to the AI, and you own the evaluation process of whether that's the right way to do it.

Shobhit: I love that you began by saying the fundamental job hasn't changed, and I love the gratitude point — I see a lot of fear and panic that everything's changed and everything people have done is irrelevant. But if you believe you learn quickly, change is an advantage, not a disadvantage.

Mick: Totally agree. What are some things that have changed — what role should product managers play more of that they currently aren't? I'll start with the most basic, obvious feedback: eval-driven development. What I've learned at Scale, and seen at Sierra, Decagon, and others — before you build an agent or a tool, you have to build your eval set or scorecard. Without it, you don't know if your training is making things better or worse. And yet this is very infrequently done — even large consulting organizations, PwC, Palantir, Sierra, Decagon, Scale, we don't always pursue eval-driven development. I think that's the number one thing to do, because otherwise you can't show progress or know which way you're trending. It's important for PMs to be plugged into that eval data generation process and make sure the eval set aligns with their product vision. A smaller change: PMs should always be dogfooding their own product and competitors' products — even more important now, because when you discover what someone else is doing, your product development time has gotten so much shorter that the ROI is very high.

Shobhit: I love that — it's further refinement of the basics. Dogfooding is such a basic thing, but you actually need to do more of it now. You're personally testing competitive coding queues, throughput incentives at Scale — what does that teach you about how PMs should work with AI systems, not just build them?

Mick: One experience early at Scale was being thrown onto a competitive coding queue called Aider — there's a benchmark with hard math problems most models at the time couldn't solve. It measures pass-at-one, and if it fails, gives the model a second chance with the output from the first — pass-at-two. I signed up as a contributor myself and solved problems, even though I hadn't coded in a decade, because vibe coding made it so much easier — I'd tell the computer "you made this mistake, go fix that" without writing a single line of code myself, and it let me empathize with my customer, the contributor. Getting back to your question — Reforge has conveyed this well — you have to learn by doing. If you want to learn by looking or reading, it's not going to get you there, for two reasons: the capabilities are changing day by day, and it's only by using them that you understand where the frontier is today. And because things change so quickly, you can't see that far ahead — it's fog of war. By doing, you learn more, and that informs what you do next — the loop starts with doing. It's even more important to be action-oriented than theoretical. Second, dog food, dog food, dog food — your product is now non-deterministic, it used to be deterministic, so you have no idea what tiny change will break it, but if you figure that out you've saved your end user a lot of time. Third, every time you try to do something, start it with AI or at least end it with AI, and put it in that process — that teaches you how to use it effectively next time. I was using AI to think about what to say on this podcast and it worked pretty well. On the flip side, I tried to use AI to write my self-review at Scale — spent three hours giving it access to every company doc, competency framework, all my PRDs and tickets, and it didn't do a great job. Took five or six hours of having the AI interview me to get to something outcome-driven and quantitative, better than I might have written myself. With everything I do, I try to embed AI in the task — sometimes it one-shots and saves me two hours, sometimes it takes two hours longer, but those are trade-offs I'm willing to live with to learn what I need to learn.

Shobhit: Last AI question — zooming out, what trends are you most bullish on that will affect PMs everywhere over the next five years?

Mick: Five years is a very long time horizon in this space, hard to actually predict. In the next 18 months: eval-driven development is going to become the norm, and it's great for PMs to be at the forefront of that rather than the tail end. Second, 2026 is supposed to be the year of agents — we thought 2025 would be, it didn't happen, but it probably will in 2026. One thing to keep an eye on: agents are very good at one-hour tasks, but long-horizon tasks that take a day, two days, three days are much harder — and every seven months or so, the time horizon it can handle is doubling. Right now AI can do things that would take a month; in two years, that's going to be two, four, eight times that. It's going to sneak up on all of us. There's also a lot happening with multimodality — I was using Gemini to update slides, it did an amazing job making them more visual, though it creates slides as images I can't edit, still valuable. Given how compressed development time is now, engineering teams can test more variance of things because it's so easy to build, and try different approaches. One thing I struggle to forecast is where AI takes over versus where human touch stays valued — 5% of all listening time on Spotify is AI-generated music, and there are apps where people intentionally listen to AI-generated music, which surprised me. I think roles like caregiver or therapist might stay more human-capable than AI-capable, but that's the million-dollar question everyone's asking: what are humans uniquely good at, what are computers uniquely good at. For the foreseeable future you're probably looking at human-in-the-loop hybrid solutions for a lot of problems. Google and Meta are already highly personalizing content and creative on the marketing and ad side — I imagine eventually that level of personalization comes into the product itself too.

Shobhit: Let's talk about creating time for innovation and creativity — PMs often say they're too busy to even think, let alone build side projects. You've founded multiple startups while leading high-stakes product orgs. How do you do that?

Mick: First, I want to acknowledge being a PM can be a very busy job — but no matter how much work you do, I can list ten more things you could do, it never ends, and you have to decide where you put the cut line. It takes a lot of time in meetings, probably the most exhausting part of my job, so I try to make sure every day I have at least two hours of non-meeting time just to stay sane. If I use my mind five days a week just for meetings, I'm not in a good state to do other things. I get that it's hard for PMs managing multiple teams with different stakeholders — driving alignment, writing documents takes time, that's real. But if you want to be successful, focus on your health first, that's the foundation for everything else — good sleep, a good mental state. Then start as small as possible — don't try to build a brand-new product over a weekend, that gets overwhelming. If you have a friend willing to join you vibe coding, it's a great collaborative experience. Every PM should be dogfooding all their competitors' products — if you're in fintech, you should have thirty different debit cards and be playing with them. Then insert AI into that process — I'll ask AI to do a competitive landscape review and compare it to my actual experience, and now I know what it does well and what I can trust it with next time. Then find side projects you'd actually enjoy using — even if you don't finish or ship it, you still learned something. Any time I spend outside producing core execution artifacts for my job, I want some of it aligned with my interests and curiosities. Take care of your health, don't feel guilty making time for this, and start small, following your interest, without worrying about being productive in the short term.

Shobhit: For somebody brand new with an hour or two a week — what should they start with?

Mick: An hour to two hours a week is plenty to get started. Three steps: this week, go watch two YouTube videos of someone vibe coding something you find interesting — sit at your computer, or put it in your headset and go on a walk. Next week, spend an hour to an hour and a half actually emulating one of those videos and building the process. After that, spend your time brainstorming and breaking your idea into chunks — two hours a week is enough to get there. The key is start simple, and there's no shortage of content on YouTube — just search "vibe coding app idea using AI" and pick a highly-rated video.

Shobhit: Good, start small, two hours a week, anybody can make that. You're building tools that rewrite news in your tone, automate gifting, and more — how do these projects help you as a product manager?

Mick: Sometimes they help, sometimes they don't, and I often don't know which case it is when I'm investing in it — it's serendipity. I love that your podcast is called Intentional Product Manager, because let's say you're working fifty-five hours a week, you're probably very intentional about fifty of those hours — these meetings get done, these PRDs get written, these decisions get made, and you work top-down that checklist. That's great, but exhausting. If you want to stay motivated and this is your side project and learning, it needs to be a little fun — don't always make it goal-oriented. The difference between what I like to do and what I have to do — going to the gym is work, going on a hike is fun. Follow your interest, don't worry about short-term gains — the dots are only obvious in hindsight, I always benefit years or months later from things I didn't know would matter. Create space for curiosity and passion and enjoy yourself, that should be your primary goal. For example, I have people in my life who are great at gift-giving, and I wanted to reciprocate, so I built an automated gifting system — here's the occasion, here's details about my sister, recommend a gift, and once I choose one, automatically order it with gift packaging and a personalized note. The aspirational version is videotaping her receiving the gift to feed feedback into the model. Similar on the news front — I built something to make sure I'm learning about topics and aware of the bias different news sources introduce. Neither of these is a billion-dollar business — they were just tools to help me.

Shobhit: Love how practical they are — here's a problem I have, I'll go solve it. Let's talk about stakeholder alignment — you have a unique take that it's deeply related to vision. How did you come to that?

Mick: Most PMs spend a lot of time in stakeholder alignment — one-on-ones well documented because you have so many stakeholders, being responsive on Slack, following up. It's a full-time job by itself. One thing I think about: the PM doesn't need to come up with all the great ideas, as long as you create a truth-seeking team where people feel comfortable contributing ideas and you create an environment to pick the best one — you've already added a lot of value. Something people don't appreciate enough: how motivating a grand vision can be to a team. Sam Altman talks about how it's easier to get someone to solve a massively challenging problem than a medium-hard one, because you're motivated by the difficulty itself. If I can paint an end-state vision — our users are very happy, we'll make a lot of money — I'd hope the engineers on my team are motivated by that vision, and even working on something small, they understand how it contributes to the bigger picture instead of feeling like they're doing backfill scripts. That's where a clear vision drives alignment — if you have multiple teams, hopefully they're building in the same direction, and it drives clarity and motivation more than anything else. I encourage every PM — yes, write your PRDs, keep the team unblocked — but if you can make time to author a good one-to-two-page vision and circulate it with the team, that's very beneficial. An example from Scale: we sell data to model makers, produced by contributors doing tasks and reviewers reviewing them. One problem we tried to solve: how do we create a world where contributors entering data don't feel like the job is lonely, and feel like they're getting better at it over time. We created a vision around investing in community, highlighting contributors who do great work, and incentivizing reviewers to send constructive feedback — we built an agent that drafts the feedback, and the reviewer just tweaks it, reducing the friction of writing constructive feedback. We started with how we wanted the user to feel, then found the features that aligned — a top-down way of thinking, rather than bottoms-up "here's some cool technology, what can I do with it." You want to do both, but the top-down part is something a PM has to own.

Shobhit: Reflecting on your career — what's a mistake you made as a product manager that you now see other people make?

Mick: That list is very long, let's do the top five. First, pay attention to your health — without it you won't work effectively, your relationships get hurt, that should be table stakes. The next level: create boundaries at work that make sense for you. I used to want to be the first person to respond on Slack no matter when it showed up, wanted to be responsive and agile — it didn't serve me well at all. As a PM it's great to have focused time to think without distractions; your job is to create a plan to get to the end state, not run a chaotic all-hands-on-deck experience. Be comfortable setting boundaries, figure out what distractions aren't serving you, and if your manager doesn't help you set those, they're not actually helping you succeed. Second: not writing enough. Writing is how you scale information within the organization and have impact — you can talk to five people a day, but two hundred people can read your documents. Use documents both to communicate ideas and for thinking — write your logic down and iterate on it, much more effective than keeping it all in your head. Third: speed isn't everything — Shreyas has a good framework where everything you do falls into one of three buckets: leverage, neutral, or overhead. If it's leverage, make time — write the PRD, the vision doc. If it's neutral, get it done quickly. If it's overhead, delegate it or don't do it at all. Fourth: as a PM, be comfortable pushing back, be comfortable being wrong — you're not going to be a valuable contributor if you only stay in your room of competency. Feel free to throw hypotheses out there and refine them with data — that applies to a lot of roles, not just PM.

Shobhit: Now the flip side — you're coaching and mentoring PMs. What separates someone good from someone undeniably great?

Mick: I'll answer, but I want your take after — you've talked to a lot of PMs in interviews and beyond. The first thing I tell mentees: simplify things as much as possible. A lot of us feel like something complicated proves we're smart, or that everything needs to be proven with data, which is nice but not always possible. Feel free to simplify, tell a very simple story of your logic, get people to buy into it — if there's data analysis to do, go do it, but as a PM you get a lot more credit for simplifying things people can understand than solving a complex problem in a complex way. Second: boundaries again, create thinking time — you won't have time to write a great doc if you're always responding to Slack. Third — and you've probably seen this in AI topics — now that AI is so good at building products, PMs need to invest in taste and judgment. The two things that get you there: spend as much time with users as possible, there's no amount of time with users that's too much; and build judgment by using competitor products, understanding what they do well and don't, and how you differentiate. Last: stakeholder management tied to product thinking — you have a vision, a roadmap, make sure that roadmap aligns with the incentives of the people who need to contribute to it, and if it doesn't, find ways to create those incentives. What have you found to be the major differentiators, Shobhit?

Shobhit: I'll try not to repeat and add some others. One: they're very persistent communicators — they learn to say the same thing over and over in different ways, because by the time other people get it, they're bored of it, and if they get bored too soon, the message never seeps into the organization. Two: they're patient for change, especially in larger organizations — I had a manager who was amazing at pushing "we need to be more data-driven," and she stuck with that change for a year until it finally happened, took a lot of effort. Three: they have a good understanding of how people really behave versus how they say they'll behave — important both from a user perspective and a stakeholder, engineer, teammate perspective. They're curious about what people actually do versus what they say they'll do.

Mick: That's very interesting.

Shobhit: Last main question, then a couple fun ones. If somebody wants to break into GenAI, what's a skill they should master right now?

Mick: A couple of angles. First, before you break in, make sure this is actually something you want to do, because breaking in takes a lot of effort — I spent a lot of time in business school trying to break into consulting and investment banking, hated both, because I'm a builder, so I came back to this. Anything you can do to simulate what the job is like helps you answer whether you actually want it, and then how to break in. On breaking in — I'd love your view too, but in my mind interviewing well and being good at the job are two very different things. Talking about getting into the interview process: it's a competitive market, PM is always competitive, AI PM even more so, but there are a lot of roles — OpenAI, Anthropic are hiring in large volumes, and there are thousands of AI apps you could work at. I think technology has broken LinkedIn and online dating in the same way — if a recruiter opens a new role, there will be five thousand resumes in three minutes, because everyone's written agents to apply for them. The whole drop-a-resume process is basically useless now, and most recruiters are reaching out directly to people. So: always find a human connection into the company, whether it's a friend of a friend who can refer you, or someone in HR — find a way in front of that line, dropping your resume won't help. Second, when you talk to people, you want to come across as genuinely excited by the opportunity, and you can't fake that — it comes from actually playing with the capabilities of the tools. Third, if you've made time to build your own tools and side projects, you have relevant stories and frustrations with current tools to share, which becomes an organic dialogue you can build on. I'd assume interviewing for a PM role at an AI company isn't that different from a traditional PM interview — still business sense, product sense, analytics — none of that really changes. But yes, would love your thoughts, since it's very hard for people to break into product, especially AI product.

Shobhit: Picking up on what you said — LinkedIn's broken, the online application system's broken, because there's too little friction on the supply side now. There used to be friction, now there isn't. It's the advice I give even on marketing a business — everyone's generating content from AI and posting it everywhere, so how do you be more human? You zig when everyone else is zagging. What can you do that makes you come across as a genuinely competitive human during the process — direct outreach, building something small, pitching differently — that's the big thing. Once you're through the interview process, all the basics have to be right, but the hard part is actually getting the interview.

Mick: Totally agree, just adding on top of that.

Shobhit: This has been amazing. A couple fun questions to wrap up. What's the last book you read and loved?

Mick: I'll give you three. If you're focused on your career, how to align stakeholders, how to get recognition without being boastful — I find Rise to be a very good book on managing yourself and your brand authentically. Second, They Called Us Exceptional, by an Indian author, about how Indians are a "model minority" and the challenges that come with facing that — a very unique book. And since a lot of my reading right now is sci-fi, to be unproductive on purpose — Three Body Problem is probably my favorite series of all time, and Children of Time, also a trilogy, highly recommend.

Shobhit: If you weren't in product, what would you be doing?

Mick: My dream job, if anyone would pay me for it, would either be a cricket podcaster — I love watching cricket, not the fastest-paced game, I get why someone might not love it, but I grew up watching it. There's a guy named Kartheek Kimber who's the stats guy, does moneyball for cricket, all the teams reach out to him, he has his own podcast — that would be a super cool job, intellectually challenging and meets my passions, I'd get to meet the players. If they wouldn't let me in there, I'd want to be a DJ — I love electronic dance music, tech house, some of my favorite music right now. I don't know anything about DJing, but it seems like a very enjoyable job.

Shobhit: Last question — where can people learn from you, follow you, connect with you?

Mick: DM me on X, or feel free to email me — I read and respond to all my emails. Happy to chat async or find time together. I'd love to meet more of your listeners.

Shobhit: We'll put that in the show notes so people have easy access. Thank you again, this was super valuable, I learned a lot — I know this is going to be a highly replayed episode to catch all the different gems you dropped here.

Mick: Thank you for having me, happy to be back anytime.

And to everyone listening — remember, product leadership isn't just about execution, it's about vision, influence, and being intentional. If you liked this episode, do me a favor and share it with somebody who'll benefit from it. Thank you again, Mick, and all of you — stay intentional. Be sure to check out our website at intentionalproductmanager.com to see how you can level up in your career.