I remember grabbing coffee with a founder I’d invested in. She was brilliant, a force of nature in sales and marketing, but she looked defeated. "Sahin," she said, her voice barely above a whisper, "I feel like I’m drowning. Everyone’s talking about AI, and I’m just nodding along. I don’t know how to code, I don’t know what a large language model is, and I feel like I’m going to run this company into the ground if I don’t figure it out."
This wasn't a first-time founder. This was a seasoned operator on her second company, with a track record of success. But the wave of AI hype had completely shaken her confidence. She felt like the rules of the game had changed overnight, and she hadn't been given the new rulebook.
I get it. The pressure is immense. You see competitors launching AI-powered features, you read headlines about multi-billion dollar funding rounds for AI startups (some of which I’ve been fortunate to invest in, like Anthropic and OpenAI), and you feel like you’re already hopelessly behind. The good news? You don’t need to be a tech wizard to lead an AI transformation. I’ve seen it happen, and I’ve helped make it happen, more times than I can count. It’s not about knowing the code; it’s about knowing the questions to ask and the framework to build.
You Don't Need to Code, You Need to Understand
Let’s get one thing straight: your job as a leader is not to build the AI model. Your job is to understand the business problem so deeply that you can guide the technical experts to the right solution. I once worked with a portfolio company in the e-commerce space. Their CEO couldn’t tell you the difference between Python and a python. But he knew his customer’s pain points inside and out. He knew that the biggest drop-off in their sales funnel was product discovery. Customers just couldn’t find what they wanted in their massive catalog of over 50,000 items.
He didn’t say, "We need a recommendation engine using collaborative filtering." He said, "How can we make it feel like a personal shopper is guiding our customers? When they land on our site, I want them to feel like we know them." That’s the reframe. He focused on the what, not the how. The tech team can handle the how. But they can’t read your mind, and they can’t read the market like you can.
Think of it like building a house. You don’t need to know how to pour the foundation or frame the walls. But you sure as hell need to know how many bedrooms you want, where the kitchen should be, and what you want the view from the living room to look like. You are the architect of the experience. The engineers are the builders. If you give them a vague blueprint, you'll get a bizarre house. If you give them a detailed, thoughtful vision of how people will live in the space, they will build you a masterpiece.
At MovieLaLa, my second company, we faced a similar challenge. We had a huge database of movies, but users were getting lost. They'd open the app, scroll for a minute, and leave. We didn't need a better algorithm, we needed a better question. The question wasn't "How do we show them more movies?" It was "How do we help them decide what to watch in the next 10 minutes?" That led to features like personalized trailers and social recommendations, which were driven by data, but conceived from a place of deep user empathy. That focus is what ultimately led to our acquisition by Gfycat.
The Three Questions Every Leader Must Ask
When I’m advising a company on their first AI project, I tell them to forget the technology for a minute and focus on three simple, brutal questions.
What problem are we really trying to solve? Be specific. "Increase efficiency" is not a problem; it’s a wish. "Our customer support team spends 20 hours a week answering the same five questions" is a problem. A real, tangible, and measurable problem. I saw a B2B SaaS company spend a year and over $2 million trying to build an AI to "optimize sales." It went nowhere. Why? Because "optimizing sales" isn't a problem. A better problem would have been: "Our sales reps spend 10 hours a week manually researching leads. How can we cut that to two hours?" That's a target you can hit.
What data do we have, and is it any good? AI is fueled by data. If you have garbage data, you’ll get garbage results. It's the most boring, unglamorous part of AI, and it's the most important. I’ve seen companies spend millions on AI projects only to realize their data was a complete mess—inconsistent, incomplete, and siloed across a dozen different systems. Be honest with yourself. Do you have the data you need? Is it clean? Is it accessible? If not, your first project isn’t an AI project; it’s a data project. Budget for it. A good rule of thumb is that 80% of the work in any AI project is data preparation. It's a painful, expensive truth, but ignoring it is fatal.
How will we measure success? The tech team might be obsessed with model accuracy, but you need to be obsessed with business impact. Who cares if your model is 99% accurate if it doesn’t move the needle on revenue, cut costs, or improve customer satisfaction? Define your key performance indicators (KPIs) before you write a single line of code. For my e-commerce company, the metric wasn’t the accuracy of the recommendations; it was the conversion rate from a product view to a purchase, the average order value, and the customer lifetime value. We tracked those numbers relentlessly. The AI was a success not when the engineers said it was, but when those business metrics started to climb.
Your First AI Hire Shouldn't Be a Data Scientist
This might be a hot take, but I stand by it. Your first AI-related hire shouldn’t be a PhD in machine learning. It should be a translator. Someone who can speak the language of business and the language of technology. This person is worth their weight in gold. They can sit in your executive meetings and understand the strategic goals, and then walk over to the engineering team and translate those goals into technical requirements. This is the person who can challenge the business team on the vagueness of their requests and challenge the tech team on the feasibility of their timelines.
I call this role the "AI Product Manager," though that role is becoming more common. Think of this person as a product manager with a deep, intuitive understanding of data. They’ve probably been a software engineer or a data analyst in a past life, but they realized they were more passionate about solving the business problem than perfecting the algorithm. They are your bridge between the possible and the profitable.
A project at a logistics company I invested in failed spectacularly because this role didn’t exist. The business leaders kept talking about "synergy" and "optimization," and the data scientists kept talking about "hyperparameter tuning" and "gradient descent." They were speaking two different languages. The project went off the rails, millions were wasted, and everyone was frustrated. Don’t make that mistake. Find your translator.
Small Wins, Big Momentum
Finally, don’t try to boil the ocean. Your first AI project should be small, manageable, and have a high probability of success. I’m a huge believer in the power of momentum. At RemoteTeam, before we were acquired by Gusto, we didn’t try to build a fully autonomous, all-knowing HR AI. We started with a simple tool that automated the process of generating compliance documents for different countries. It wasn’t glamorous, but it solved a huge pain point for our customers. It worked.
That small win did two things. First, it saved our customers a ton of time and legal headaches, which they loved. Second, it showed our own team that this AI thing was real and could actually help us build a better product. It gave us the credibility and the buy-in to tackle bigger and more complex projects. It created a virtuous cycle of success. Our next project was an AI-powered payroll calculator that could handle the complex tax laws of different countries. We never would have had the confidence or the resources to tackle that if we hadn't gotten the first win with the compliance documents.
So, what’s your small win? What’s the one, simple AI project you can launch in the next 90 days that will deliver real, measurable value? Don't aim for a moonshot. Aim for a stepping stone. The moon will still be there later.
Leading an AI transformation is not a technical challenge. It’s a leadership challenge. It’s about having the clarity of vision to identify the right problems, the courage to ask the tough questions, and the humility to start small. The leaders who win with AI won’t be the ones who can code the best algorithms. They’ll be the ones who can ask the best questions and build the best teams to answer them. Stop trying to be a tech expert and start being the expert on your business. That's the only expertise you need.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.