What They Don't Tell You About Entrepreneurship in the Age of AI
Unlock the secret to startup survival in the AI era. Discover why most products fail, how to build unshakeable customer trust, and why your "growth muscle" is your biggest asset.
Last week, I was invited by the AI and Technology Academy, a joint initiative of Google Turkey, Turkey’s Ministry of Industry and Technology and the Turkish Entrepreneurship Foundation, to speak with a group of selected young talents from across the country.
The topic was: “Making a Difference in the Age of AI”
For an hour, they asked genuinely good questions. Not surface-level ones, they wanted to go deep.
I wanted to put everything we talked about into writing. As I was speaking, I realized there are things I don’t say out loud very often. The questions had to pull them out of me.
So let’s start from the beginning.
THIS ISSUE IS BROUGHT TO YOU BY OUTSKILL
Still juggling 10 tools? Replace 80% with AI workflows
Perplexity’s new AI can think, design, code, and manage projects in one place, boosting productivity 10x.
Join Outskill’s 2-day LIVE AI Mastermind to build automations, create agents, and gain a competitive edge.
Free for 72 hours + bonuses included.
Failure Is the Work Itself
When they introduced me; “20 years, three ventures, 50+ products…” I opened with: “Most of them failed”.
People are sometimes surprised when I say that. But we really shouldn’t be anymore.
We all know Michael Jordan as a legend. The highest scorer, the player who showed up in the biggest moments, the undisputed best of his era. But that same Michael Jordan also ranks among the all-time leaders in missed shots in NBA history. We almost never talk about that.
Entrepreneurship works exactly the same way. The successes you see from the outside are built on top of dozens of failures happening on the inside.
Honestly, it’s hard to call someone an entrepreneur if they quit after the third or fourth failure. Accepting this is no longer difficult for me.
But accepting it is one thing. Analyzing it is another.
How Do You Save a Failing Product?
When a product fails, the biggest mistake is panicking.
The second biggest mistake is adding new features.
Every product is a theory. The theory that “these people will buy this solution to this problem at this price”. Failing means the theory didn’t hold. So the first order of business is asking:
Did I actually test this theory correctly?
Most of the time, the answer is no. Either the testing wasn’t done, or the wrong things were measured.
Measuring means identifying where the drop-offs are.
Say you haven’t made a single sale. Think backwards: 1 sale requires 10 registered users. 10 registrations require 100 visitors. 100 visitors require 10,000 clicks. 10,000 clicks require reaching 1 million people.
Now ask yourself honestly: Did you actually reach that million?
Most of the time, no. You reached 200 people and wondered why there were no sales.
If the reach is sufficient, the second step is: Where did I perform worst in the funnel?
Are people bouncing from the site immediately? Are they arriving but not registering? Registering but not buying? Each scenario has a different fix.
The third step is looking at the business model. Is the price wrong, or is the model itself broken? Should it have been a subscription? A one-time purchase?
Don’t skip that order. Most people jump straight to pricing. But changing the price before testing reach and conversion is like fixing the roof before painting the walls.
The Big Idea Trap
“How do we know if an idea is big?” they asked.
The honest answer: that question itself is a little off.
Investors always say the same thing: “Find a business that can 100x”. But achieving 100x isn’t just about intelligence or talent. It depends on capital, timing, environment, luck, and a thousand other variables.
So set aside what investors want.
One of the biggest obstacles in entrepreneurship is trying to think too big from the start.
The probability that any of us — myself included — will build something truly massive is genuinely very low. There are thousands of obstacles standing between you and that dream.
Out of millions of startups, only a handful reach that scale. It’s no different from buying a lottery ticket. And investors, at their most basic, are simply in the business of buying lottery tickets.
So especially if you’re just starting out, instead of chasing a “big” idea, focus on finding something small but indispensable.
It doesn’t have to save lives. It can be something fun, something mildly addictive, something small. But it should be tangible and testable.
The most critical thing is getting your idea in front of people. And you don’t need to go far for that: an uncle, an aunt, someone you know from the business world.
Start with people who can evaluate your idea honestly.
You don’t need to be big on your first venture. Experience accumulates, networks grow. By your tenth venture, you can aim for “big”.
Right now, the only thing that matters is increasing the number of people your idea can reach and actually being able to test it.
Winning Customers?
“What’s the most unexpected path to user satisfaction?” they asked.
My answer was two words: sincerity and transparency.
Everyone can deliver quality now. Product standards have risen. Being excellent is the minimum expectation. It’s no longer a differentiator.
What differentiates you is being genuinely on the customer’s side.
The most unexpected yet most powerful move is this: sitting in the customer’s chair and honestly telling them that your product might not fully solve their problem.
I know it sounds counterintuitive. But that’s exactly why it works. Because no one else does it.
When everyone else is trying to sell their product and you say “actually, this other thing might be a better fit for you”. That person never forgets you.
You might lose the sale in that moment, but you gain something that can last a lifetime: trust. The kind that keeps you top of mind when a future need arises.
Our most loyal corporate clients were built through exactly this kind of relationship. At a certain point, I told them: “We might not be the right product for you”. Because at the time, it was true. That customer stopped seeing me as a sales partner and started seeing me as a solution partner. The difference between those two things is enormous.
If you want to build something that lasts, leading with “let’s just make the sale” shortens that road considerably.
Going Global?
“What’s the biggest obstacle to global growth for digital products?”
Not technical. Not marketing. Not pricing.
Culture.
Taking a product into another market isn’t just about translating the language. It’s about understanding what makes those people laugh, what bores them, what they react to instinctively.
Without that understanding, every ad, every message, every campaign, no matter how well designed, lands like a foreigner speaking with an accent. It’s understood, but not felt.
We had a product that let you send a fax from your phone. Japan became an unexpectedly large market for us. We couldn’t figure out why, so we investigated.
It turned out that in Japan, sending an email is culturally considered disrespectful in certain contexts. A fax, on the other hand, is a sign of great respect. If we’d tried to market in that country without knowing that, nothing would have stuck.
The more countries you expand into, the more cultural codes you have to crack. It’s not something to take lightly. But because most of your competitors skip this part, there’s a real opportunity hidden here.
Where Is Your Customer?
“What’s the most underrated yet most effective method for global growth?”
Understanding the daily life of your target audience.
When they wake up in the morning, where do they look first? What platform do they spend time on? What communities are they part of?
That’s what advertising actually is. If you want to reach 1 million people but don’t know where they are, you’re searching in the wrong places.
If you’re selling a B2B product, the executive you’re targeting might be watching cat videos right now. In that moment, they’re not ready to consume your message. Finding the place where they are ready that comes before everything else.
There’s no single formula for this. But the most effective strategy is simply being visible where your customers already are.
Not big-budget campaigns. The right place, at the right moment.
Who Wins: PMs or Developers?
“How is the balance between product managers and developers shifting in the age of AI?”
I know you’re expecting me to say “PMs will write code and developers will be out of a job”. I think differently.
Both roles will weaken.
Here’s the logic: if production is no longer the challenge, managing production stops being the challenge too. If you can build most things with a few prompts, the real question isn’t “how do I build this?”, it’s “how do I grow this?”
I think we’ll slowly see these two roles melt into a single thing: growth muscle.
Everyone will have to become a growth operator of some kind. And just like software development eventually branched into backend, frontend, DevOps, and so on, growth will become a deeply layered area of specialization in its own right.
The winner is whoever understands growth. Whoever understands the customer. Whoever knows how to sell to whom. Whoever solves distribution. Regardless of title.
The Era Where Everyone Needs to Know a Little of Everything
“Doesn’t knowing a bit of everything spread you too thin? Isn’t specialization the most important thing?” they asked.
Not long ago, the answer to that would have been yes.
Expertise really did pay. There were few people who knew certain things, and they could monetize that knowledge.
Not anymore.
When you ask the right question, AI walks you through it step by step. Knowing everything now effectively means being able to do everything.
We’re a team of seven at Next Big App. Whatever the team does and whatever it can’t do, ultimately falls on me. Software, marketing, sales, finance, operations… I have to understand all of it well enough to engage with it.
And this is a structural shift. The era of 50–60 person companies is probably ending. Margins are compressing, competition is intensifying. Small, versatile, fast-moving teams will come out ahead. On a team like that, being informed across every domain isn’t an advantage anymore. It’s a requirement.
We’re entering a period where we need to question every team member’s direct contribution to growth.
What Separates a Good Product from a Bad One?
“In a world where AI has accelerated prototyping this much, what will separate a good product from a bad one?”
One word: trust.
If everyone can build everything, the market becomes flooded with products. When speed is no longer an advantage, the only thing left is: trust in that product, trust in the person behind it.
An executive at a16z framed it differently: software will become like content. Consumable. Use it, discard it. In that world, the founder has to become the brand.
BYD outperforms Tesla in certain efficiency metrics. But we talk about Elon Musk. We follow the person, not the product.
Every founder needs to build their own media. A newsletter, YouTube, talks, writing, whatever fits you. Because as products start to look more and more alike, people go looking for a name they can follow.
Does Automation Alienate Users?
“Can too much AI, too much automation, push users away?” they asked.
It can. But the problem isn’t automation. It’s the loss of the feeling of impact.
People don’t care about the outcome as much as they care about how they contributed to it. Misunderstanding this is a serious mistake.
When boxed cake mixes first came out, they had everything inside. Just add water, bake, done. But they didn’t sell.
Then the packaging was changed to say: “You add the egg”. Same product. This time it sold. Because for the person baking, the feeling of “I made something too” kicked in.
Most crosswalk buttons are non-functional. They’re not electronically connected to the traffic lights. They’re just there. When people press them, they feel slightly less helpless. They get to say “it stopped because of me”.
The best digital example is flight search sites. Technically, they could return results in under a second. But they make you wait 30–40 seconds. Why? To send the message: “I’m working really hard for you”. Perception overrides reality.
The point where you should stop with automation is exactly here: the moment where the user still feels like they’re touching something, influencing something.
When AI Optimizes Everything, What Happens to Creativity?
This was one of the questions that stayed with me for a while.
Let me reframe it first. AI doesn’t just optimize, it can be creative too. You say “make an unexpected connection between these two unrelated things” and it produces something genuinely creative.
So the real question isn’t about AI’s creativity. It’s: what do we do with our own creativity from here on out?
I think the space for being creative in business, for solving problems creatively has largely been closed off. We don’t need to spend energy there anymore.
So what’s left?
Art.
Creativity fused with human emotion. I recently visited an exhibition called Folia at the Abdülmecid Efendi Pavilion in Istanbul, and it was extraordinary. What I felt in that moment is hard to describe, something like belonging to a vast universe. AI cannot do that. It won’t.
Maybe this isn’t a loss. Maybe it’s a freedom. With the pressure of business creativity lifted, we’ll have more time to look inward.
The Era of Permanent Products Is Over
They asked what my greatest fear is when designing an AI product.
Shelf life.
In the past, it would take a competitor five to ten years to surpass you. Now, if you’re lucky, it’s measured in weeks.
You see it constantly. Google, OpenAI, Claude… A model that’s considered excellent gets the comment “this is already outdated, there’s something better now” two weeks later.
The only way to live with that fear is to think in terms of multiple products. The advice VCs have repeated for years, “focus on one product” is no longer valid. It’s outdated. Frankly, I think it was bad guidance to begin with.
Tying everything to a single product is like investing all your money in a building with no foundation.
The right mindset is this: own a market. And be able to think about dozens of products within that market simultaneously.
Focus on the people you’ve already reached. Someone you connected with on LinkedIn, someone you met five years ago, a former colleague. Trust starts there. People will buy your products because they know you, not because you have no competition.
The Most Overhyped and The Most Underrated?
“What are the most overhyped and most underrated trends of the next five years in AI?” they asked.
The most overhyped: intelligence itself.
There’s a body of research showing no correlation between success and IQ, in fact, sometimes a negative one. Smart people aren’t always more successful. So why are we talking as if AI will rule the world simply because it’s intelligent?
Current models already score above 120–130 on IQ tests. That question is settled. The things intelligence still can’t reach are too numerous to count.
The most underrated trend: the disappearance of the interface.
There was the keyboard. Then the screen, then touchscreens. Every time, there was some intermediary between us and the software. Now we’re moving to managing systems through voice. Very soon, through thought and intent.
This is the biggest UX leap in the history of software. And almost no one is talking about it.
Where it’s all heading is even more interesting: a time when most needs are met simply by thinking, where perhaps no one is “selling” anything to anyone. What that does to our social lives is something I genuinely wonder about.
The Hook: Speaking the Language of Someone’s Pain
“What hooks a customer in the very first interaction?” they asked.
Every day, someone receives dozens of emails, dozens of messages. Why would they read yours?
One answer: speaking the language of their problem.
Almost every good SaaS landing page opens with the same structure: “You have a problem like this, don’t you?”
Here’s an example from our own product: “You spend all day on the phone calling thousands of people. And you know: 900 of them are going to hang up. Do you want to keep living with that pain — or would you rather have AI make those 1,000 calls for you, and bring you only the 10 people who are actually ready to buy?”
That’s the hook. Understanding that daily frustration. And making clear that you have the solution.
Thinking Outside the Rules: Listen to the People Who Say “Impossible”
“What’s the best example of unconventional thinking for AI entrepreneurs?” they asked.
Really listen to the people around you.
Most of us make a lot of assumptions. But the customer lives in a completely different world. I’m walking them through hundreds of features. And they say: “You know what, I just need it to tell me I’m in a meeting when someone calls”.
Sometimes the need lives in a very small place.
That’s why you should look at the things people say are “impossible”. But be careful: “impossible” usually doesn’t mean truly impossible. It means “no one has done this yet” or “I don’t know that it can be done”. And with AI, that equation changes.
Having simultaneous phone conversations with thousands of people, making them feel like real conversations, at a small budget that used to seem impossible. It isn’t anymore.
The impossibilities are where the best startup ideas are hiding.
How Do You Find Your First Customer With No Budget?
The answer is simple, but most people don’t want to hear it: start with the people closest to you.
Our first major enterprise customer was a reader of this newsletter. I wrote about what we were building, what the product could do. The next day, I got a message: “Can we try it?”
Zero ad spend. Zero middlemen.
It doesn’t have to be a newsletter. It can be LinkedIn, an old colleague, someone your uncle knows. Your first customers come from people who already know you, who already trust you.
If I Were Starting From Scratch, What Would I Do in the First 30 Days?
Great question. I read it as: what should we focus on?
I’d knock on 100 doors.
Message people on LinkedIn, call old colleagues, visit someone your uncle knows. Ask for thirty minutes. Listen to their problems. Build on top of what you hear.
We can come up with brilliant ideas in our own heads. But when those ideas meet the real world, people don’t find them brilliant. Sometimes they’re too sophisticated. You’ve built 20 features, and they say “that’s not my problem at all”.
If you don’t know their problem, you can’t sell them anything.
Will Face-to-Face Communication Regain Its Value?
Absolutely.
As we digitize, automate, and shift to voice-managed systems, the bond between two people in the same room will continue to be something entirely different.
We consume everything in 3 seconds now. A 3-second video, a 3-second clip of a song, we don’t even look at the photo we just took. At a Coldplay concert, we watch them through a screen while recording a video to post on Instagram. Strange times.
Face-to-face communication is nothing like any of that. And it won’t be. The gap between influencing someone in the virtual world and talking to someone right beside you isn’t closing, it’s widening.
The greatest benefit, the deepest trust, we’ll get those from the relationships we build face-to-face. That won’t change.
A lot of ground was covered in that one hour. The audience came prepared.
I hope it’s useful for those reading it here too.
Until next time,
Selim