Growth Ritual #80
📋 In This Issue:
Think Like Uber: How to Join the 5% Club of AI Winners
This AI Vision Automation Makes Your SaaS Stack 100x Cheaper — 🔒
Emad Mostaque Just Leaked the Cheat Codes to the AI Economy — 🔒
Know someone who’d love this? Invite them and you’ll both win—your friend discovers the newsletter, and you unlock 1 month of premium access.
Think Like Uber: How to Join the 5% Club of AI Winners
The MIT Sloan School of Management recently published its “State of AI in Business 2025” report. If you haven’t read it, you should.
The takeaway is as stark as it is simple:
“95% of corporate AI pilot programs fail to make a measurable impact on the company's bottom line”
The knee-jerk reaction for many is to share this headline with a dismissive take: “See? AI isn't all that important. Let's get back to our daily jobs”.
That conclusion couldn't be more wrong.
Ninety-five percent. If we only look at the headline, failure seems inevitable. But we need to dig deeper. The problem isn't that the technology is useless. The problem is that the fundamental approach is broken.
Too many companies are desperately trying to answer the question, “How can we implement the latest AI solutions?” without first asking the only question that matters.
Meanwhile, a booming industry of “AI consultants" and “AI experts” waxes poetic about the incredible future and philosophy of artificial intelligence. To managers in the trenches trying to get things done, this is just hot air.
While these futurists are busy debating long-term, science-fiction scenarios, a quiet minority of operators are focused on solving concrete, urgent problems that actually make money.
This has split the world of AI adoption into two distinct camps.
Camp 1: The Philosophers
These are the brilliant minds obsessed with the "what if" scenarios.
Mustafa Suleyman, the CEO of Microsoft AI, warns against researching AI consciousness, fearing it could lead to “psychotic crises” in humans who become too attached to chatbots. He is preoccupied with the distant and theoretical risks of emotional harm.
Researchers at Anthropic are exploring “AI welfare", believing that models like Claude should have the right to end a conversation if they feel they are being “mistreated”. They are attempting to grant agency to a machine before it has any real-world footing.
These topics make for fascinating dinner party conversations. But for an entrepreneur trying to build a business, they are, frankly, noise.
Obsessing over massive, existential problems that we have no immediate control over is far less productive than focusing on a single, pressing question: “How can I create value today?”
Camp 2: The Pragmatists
These are the people building real things in the real world. They are focused on measurable outcomes.
Uber's Self-Driving Cars. Uber's CEO isn't debating whether his autonomous vehicles have feelings. He's solving a cold, hard problem: building a safe and reliable driverless car. His approach? A multi-modal sensor fusion of cameras, LiDAR, and radar to achieve "superhuman safety levels." He uses redundant systems to solve a tangible problem. He isn't tying the company's fate to a single technology (like Tesla's camera-only approach) because the cost of failure is a human life. He's not being dogmatic; he's being pragmatic.
E-commerce Recommendation Engines (e.g., Amazon).
The Concrete Problem: When a customer views or adds a product to their cart, what other items can we show them to increase the order value? This is a direct revenue problem.
The Pragmatic Solution: No one here is trying to build an AI that "understands the customer's soul". The engine is typically powered by simple statistical models like "association rule learning" or "collaborative filtering". The system's logic is blunt: "Thousands of people who bought Product A also bought Product B. Let's show this new customer Product B". It analyzes data crumbs from millions of users to find the most probable statistical matches.
The Measurable Impact: An increase in average order value, higher conversion rates, and a direct, positive impact on revenue. This is a multi-billion dollar revenue stream built on simple logic.
Logistics and Courier Route Optimization (e.g., UPS).
The Concrete Problem: A courier has 15 different packages to deliver. What is the optimal sequence and route to minimize fuel consumption and time? This is a classic operational efficiency problem.
The Pragmatic Solution: This isn't an AI that "predicts the future". It's modern applications of a classic algorithm known as the "Traveling Salesman Problem". The system calculates the most efficient route in seconds by processing delivery addresses, real-time traffic data, and estimated service times. It uses dry, boring data to find the most efficient path.
The Measurable Impact: Lower fuel costs, more deliveries per courier per day, and increased customer satisfaction. In short: doing more with less.
Fraud Detection in Banking (e.g., Every Major Bank).
The Concrete Problem: If a credit card is used in Istanbul, and three minutes later the same card is used for an online purchase from Brazil, is it a legitimate transaction or fraud? The wrong decision either costs the bank thousands or alienates a real customer.
The Pragmatic Solution: The system doesn't try to "read a criminal's mind." It's a machine learning model trained on millions of transactions to perform "anomaly detection". It compares dozens of "boring" data points: location, time, amount, merchant type, and the user's normal spending habits. If a transaction deviates sharply from the norm and resembles previously flagged fraudulent transactions, it's instantly blocked.
The Measurable Impact: Preventing billions of dollars in fraud annually. This is a massive, direct defense of the bottom line.
The Unwritten Rule: The Uber Mindset
This brings us back to the 95% failure rate. Companies are diving headfirst into AI without understanding its technical limitations or its most suitable use cases. They are approaching AI like philosophers, not pragmatists.
The secret to success lies in the simple Uber Mindset.
Instead of asking, "How can we use AI?", they should be asking:
"What is our single most impactful business problem right now, and what is the best way to solve it?"
Notice the question isn't even "How can we solve this problem with AI?"
The solution might not be a technology at all. It could be a change in perspective, a process redesign, or a new business model. The key is to first obsess over the problem and its root causes. Only then can you identify the real obstacles and determine the right tools to overcome them.
Our goal isn't to build a conscious machine. It's to find a solution that is "good enough to get the job done".
You don’t need a multi-million dollar AI to write better marketing copy. You need a simple prompt template.
You don’t need a state-of-the-art computer vision model to automate competitor analysis. You need a Zapier integration that scrapes a few websites for you.
You don’t need a complex AGI to manage customer support. You need a well-trained, fine-tuned chatbot that can answer the 100 most frequently asked questions.
The most successful AI projects I’ve seen aren’t trying to change the world. They’re trying to save an employee 10 hours a week on a tedious task. They focus on measurable outcomes that directly impact the P&L.
So let's leave the philosophers to debate the consciousness of AI—it’s a fascinating topic for another time. But let's not turn our backs on the rising tide of technology just because 95% are failing.
If everyone today were trying to drink water through their nose and failing to quench their thirst, we wouldn't say, "Let's stop trying to drink water", would we? We’d simply find a better way to do it.
Keep reading with a 7-day free trial
Subscribe to Next Big App to keep reading this post and get 7 days of free access to the full post archives.