The Art (& Science) of finding the most important AI Use Cases - Part 1
Why Your AI Roadmap Looks Like Everyone Else’s—and How to Fix It. This is 7 part series on the art & science of discovering most important AI use cases
🔑 Key Takeaways
Common Approach: "Ordering off the menu."- Teams pick from a list of predefined, popular AI use cases—resulting in generic, copy-paste AI roadmaps packed with low-ROI initiatives and little strategic differentiation.
Avoid the AI Herd Mentality: Too many organizations blindly follow trends—adopting the same AI use cases without aligning them to their unique strengths or strategic goals. The result? Underwhelming outcomes, wasted investments, and the false conclusion that “AI is just hype.”
Embrace First-Principles Thinking: Instead of copying prevalent AI use cases, businesses should analyze their specific challenges and opportunities from the ground up to identify impactful AI use cases.
Applied AI Requires First-Principles Thinking: Successful AI products stem from deep insights into user behavior and needs, creating experiences that resonate and build love & reverence, leading to customer delight.
Originality in AI Use Cases Drives True Innovation
The transformative power of AI is realized when organizations develop unique, high-impact use cases tailored to their business context, rather than following industry trends.
Today, unfortunately, AI roadmaps of most organizations look the same.
The success of any AI-first product comes down to a few key things—and one of them is: what problem or pain point are you solving? In other words, what's your AI use case?
Unfortunately, most teams and leaders today are fixated on a single question: “Are we using the best AI model or algorithm?”
This stems from a deeply misplaced belief that using a state-of-the-art model is the guaranteed path to AI success—when in reality, success depends far more on many other factors including problem framing, data quality, and integration into business workflows etc [as showing in Fig 1]
In our AI experience spanning two decaded, one of the most critical factors for AI success is this: Are you solving the right—and most important—problems?
Even the most advanced AI won't deliver impact if it's applied to the wrong challenges. Problem selection, not model selection, often determines the difference between meaningful ROI and wasted effort.While AI holds incredible promise, the current landscape is crowded with a rinse-and-repeat approach to use cases—chatbots, content summarization, recommendation engines—recycled endlessly across industries. Most VCs, Founders, and CXOs just want a handy list of AI use cases from others rather than themselves doing the necessary thinking and encouraging likewise within their teams and organizations.
There's often little thought given to what truly differentiates a business. Which is why, regardless of an organization’s value proposition, industry, strategy, or unique strengths, the AI roadmap ends up looking eerily similar. Right now, for most companies, Fig 2 depicts how the AI roadmap for most organizations looks like:
So why is this happening? In part, it’s because of OpenAI’s staggering success. From zero to a $300B valuation in just nine years—and just last month it raised a record-breaking $40B—they’ve become the crown jewel of the AI world. Today, thanks to its cutting-edge technology, OpenAI is the company everyone wants to emulate. Startup founders, MNC CEOs, VCs, PEs—even entire countries—are chasing that same dream.
And in the process, there’s this growing belief that it’s all about the technology. That breakthrough algorithms and state-of-the-art models are what matter most. This makes complete sense… if you're building the next top industrial research lab like OpenAI, DeepMind, Hugging Face, Meta AI, or Boston Dynamics.
But if you're in the game of applied AI—which, let’s be honest, is 99.9% of us—what you’re solving matters way more than how you solve it. And that takes creativity. It takes first-principles thinking to uncover meaningful, high-impact use cases.
Unfortunately, very few leaders want to do the first principles thinking on this. Hence, most teams skip that part. Instead, they chase the usual suspects in the name of “low-hanging fruit.” The result? AI roadmaps across companies today look less like innovation… and more like the great migration through the Serengeti—everyone following the herd.
So, what’s the problem if everyone’s implementing the same AI use cases? Technically, nothing. But here’s the catch: the true transformative power of AI remains untapped.
The real magic of AI isn’t in doing what everyone else is doing
It is in implementing original, high-impact use cases grounded in first-principles thinking and a deep understanding of your business.
That’s where AI stops being a buzzword and starts becoming a game-changer.Below, I would be sharing 5 examples of such original, high-impact AI use cases—ones that stood apart from the usual playbook and are deeply rooted in business context and first-principles thinking:
Spotify’s Discover Weekly (Part 2)
Google’s finding special character (Part 3)
DHL’s needle in the haystack (Part 4)
Airwoot’s Noise Cancellation (Part 5)
Vahan’s Blue Collar hiring conversational system [How we won over Sam Altman & Vinod Khosla] (Part 6)
We will end this series with a comprehensive framework to identify potentially high ROI AI use cases. (Part 7)
Stay tuned!
If you use this material in part or in full, please do cite this write-up as:
Gupta, Anuj. (Apr 2025). The Art (& Science) of finding the most important AI Use Cases. anujgupta.com https://pragmaticai1.substack.com/p/why-your-ai-roadmap-looks-like-everyone
or
@article{gupta2025ai-use-cases-1,
title = {The Art (& Science) of finding the most important AI Use Cases -
Part 1},
author = {Gupta, Anuj},
journal = {anujgupta.com},
year = {2025},
month = {April},
url = {https://pragmaticai1.substack.com/p/why-your-ai-roadmap-looks-
like-everyone}
}