I’m going to say something that might sound crazy. Stop looking for “AI Product Managers.” They don’t exist. At least, not in the way most founders and hiring managers think they do. The job descriptions I see floating around Silicon Valley are a laundry list of impossible-to-find qualifications, basically searching for a unicorn that has a PhD in machine learning, ten years of product experience, and the design sense of Jony Ive. It’s a fantasy.
After two exits and investing in over 200 companies, including some of the foundational players in the AI space like Anthropic, OpenAI, and Scale AI, I’ve seen this mistake play out dozens of times. Founders burn months searching for a mythical candidate while their AI strategy stagnates. At RemoteTeam, before we were acquired by Gusto, we made this exact error. We hired a brilliant, traditional PM to lead our first major AI initiative. He was incredible at shipping features for our core product, but the AI project went nowhere for a year. It wasn't his fault; it was mine. I was hiring for the wrong things.
Forget the generic frameworks you read online. I’m going to share the playbook that actually works, the one based on seeing what separates the top 1% of product leaders who can actually ship AI products that millions of people use.
1. Stop Over-Indexing on “AI Experience”
This is the biggest filter that leads companies astray. You see a resume without the words “Machine Learning” or “NLP” plastered all over it, and you pass. Huge mistake. The field is moving so fast that specific experience with a model or technique from two years ago is likely already obsolete. What you need is not someone who has used AI, but someone who can think in a way that’s compatible with how AI works.
What does that mean? It means you should prioritize systems thinking and a deep, almost obsessive, curiosity about how complex systems work. When I was involved with teams at some of the biggest tech companies, the best AI PMs weren’t the ones with the computer science PhDs. They were the ones who could whiteboard the entire flow of data, from user input to model inference to the feedback loop that retrains the model. They could reason about probabilities, edge cases, and second-order effects.
One of the best hires I ever made was a woman who had been a product manager for a logistics and supply chain company. On paper? Zero AI experience. But in her interview, she explained how she optimized shipping routes by treating the entire network as a dynamic system, constantly adjusting to new inputs like weather, traffic, and fuel prices. She talked in terms of inputs, outputs, and feedback loops. I hired her on the spot. Within a year, she was leading one of the most successful AI product teams in the company.
2. They Must Be Data-Obsessed, But Not How You Think
Every PM claims to be “data-driven.” That’s table stakes. But for an AI product, it’s a different game. It’s not just about looking at dashboards and A/B test results. It’s about having an intuition for the shape of data itself.
An AI product is, in many ways, a product built out of data. The model is only as good as the data it’s trained on. A great AI PM understands this in their bones. They don’t just ask for “more data.” They ask: What data do we have? Is it clean? Is it biased? What data don’t we have that’s causing the model to fail in weird ways? They spend more time with data scientists looking at raw data distributions than they do in Figma.
Here’s a real example. A portfolio company was building an AI-powered code completion tool. The model was great, but users weren’t adopting it. The PM, who was just okay, looked at the adoption metrics and suggested marketing campaigns. The great PM they eventually brought in did something different. She spent a week shadowing engineers. She discovered the model was fantastic at completing boilerplate Python, but fell apart on the company’s specific, esoteric JavaScript framework. The problem wasn’t the users; it was a massive gap in the training data. They spent a month gathering and labeling internal code, retrained the model, and adoption shot up 400%.
That’s the difference. One looks at a spreadsheet; the other understands the soul of the machine.
3. Look for Scrappiness and a High “Figure-It-Out” Quotient
Building in AI is building on shifting sands. The models change, the platforms change, the user expectations change—all on a monthly basis. A PM who needs a perfectly defined roadmap and a stable environment will fail, guaranteed. You need someone with an insanely high “figure-it-out” quotient (what I call a FIOQ).
This is the person who, when faced with a totally new problem, doesn’t freeze. They have a process for learning. They find the five experts on the internet and email them. They download a new open-source tool and get it running on their laptop over the weekend. They build a crappy prototype using a no-code tool just to see if an idea has legs.
When you interview, don’t ask them to solve a hypothetical Google-style brain teaser. Give them a real, messy, unsolved problem your team is facing. For instance: “We want to use AI to reduce customer support tickets, but our support data is a mess of emails, chat logs, and call transcripts. What would you do in your first 90 days?”
A bad answer is: “I’d work with engineering to build a data pipeline.” It’s generic and passive.
A great answer is: “First, I’d manually read 500 of them myself to get a feel for the patterns. Then, I’d probably use a simple off-the-shelf tool like OpenAI’s API with a basic script to try and categorize the top 20 ticket types. I’d want to see if a simple model can even get 60% accuracy before we invest a single engineering cycle. I’d have a rough prototype in a week that we could use to convince leadership this is worth doing.”
See the bias for action? The scrappiness? That’s what you’re looking for.
4. The Best AI PMs are Translators
Finally, this role is fundamentally about communication. An AI PM sits between some of the most technical people in your company (the ML engineers) and the rest of the business (sales, marketing, leadership, users). They have to be a world-class translator.
They need to be able to go into a room with PhDs and have a credible conversation about precision vs. recall, and then walk into a room with the CEO and explain why the model is hallucinating and what the business impact is, all without using technical jargon. This is an incredibly rare skill.
It’s the ability to explain a confidence score not as a statistical measure, but as a business risk. It’s the ability to explain a feature not as “we’re using a transformer model,” but as “the product can now understand the intent behind what you write, not just the keywords.”
When you hire, test for this explicitly. Ask them to explain a complex AI concept to you as if you were a new salesperson. Ask them how they would handle a situation where the engineering team says a feature is impossible. The best candidates won’t just defer to the engineers; they will ask probing questions to understand the constraints and then work to redefine the problem into something that is possible.
So, my advice is simple. Burn your unicorn job description. Instead, look for a scrappy, data-intuitive systems thinker who can translate between humans and machines. They might be working at a logistics company, a finance firm, or a gaming studio. They probably don’t have “AI” in their job title today. But they are the ones who will actually build the future of your company. Find them, and you’ll be miles ahead of everyone else still searching for a fantasy.
Frequently Asked Questions
Do I need technical skills to hire your first ai product manager (and what to look for)?
Not necessarily. While technical understanding helps, the most important skills are clear thinking and the ability to break problems into smaller pieces. Many successful founders I've invested in started with zero technical background and either learned enough to be dangerous or found the right technical partner.
What tools do I need to get started?
Start with the basics. You don't need expensive software or fancy tools. A spreadsheet, a note-taking app, and direct access to your customers will get you further than any enterprise platform. Add tools only when you hit a specific bottleneck.
How do I measure success with this approach?
Pick one or two metrics that directly tie to your goal and track them weekly. Vanity metrics like page views or follower counts rarely matter. Focus on metrics that reflect real engagement or revenue impact.
How long does it take to hire your first ai product manager (and what to look for)?
The timeline varies depending on your starting point and resources. For most founders, expect 2-4 weeks for initial setup and 2-3 months to see meaningful results. I've seen teams move faster when they focus on one thing at a time rather than trying to do everything at once.