Mala Ramakrishnan, Founder and Managing Partner at Progressive Ventures, fueling the next wave of AI innovation, out of Silicon Valley.
AI wrappers have been a buzzword since startups started using ChatGPT as their unofficial co-founder. It’s never been easier to slap “AI” onto a product and call it revolutionary. But being easy to build doesn’t mean it’s easy to win.
Many startup gurus will say execution is primary and the idea is secondary.
But what if I said that 35% of startups fail due to lack of product-market fit, according to an analysis by CBInsights? They don’t see the light of a profit-generating day because the problem they’re solving doesn’t exist.
The result? A cemetery of ideas that sounded great at the investor’s doorstep but fell apart in the market. Execution is critical, but what you’re executing on matters just as much. So how do you decide whether your AI idea is worth building a company around? Let’s break it down.
Before we even touch AI specifics, let’s get some startup fundamentals straight:
1. What’s The Total Addressable Market (TAM)?
This isn’t just a pitch deck slide but also your internal litmus test as a founder for assessing how big the market is and if there will be enough people to pay for it. Do you have access to sufficient numbers of customers to build a $1 million revenue stream in the next 6-12 months?
2. Do Your Unit Economics Make Sense?
Are your numbers sustainable?
• Customer Acquisition Cost (CAC): If your CAC is $10, but your customer only brings in $6, you’re in trouble.
• Gross Margin: AI startups often have high cloud or GPU costs, so understanding gross margin is vital to scalability. If you make $10 per deal on your customer, but have to spend $20 per customer on LLM overhead costs—especially after your credits run out—your gross margin is in the negative, and you don’t have a product that can sustain itself.
• Net Retention: For every one customer you gain, are you losing two to an incumbent who is modernizing their stack and giving their product to your customers for free?
3. What’s Your Edge?
How does your idea compare to what’s already out there? Are you creating a 10-times-better solution, or will you drown in a sea of lookalikes?
Now, there’s a ton of information on SaaS product-market fit out there.
But How Does This Change In The World Of AI?
AI isn’t new. What’s changed is how accessible it’s become. Foundational models like GPT have lowered the entry barriers, meaning the bar for great ideas is now sky-high. Here’s how to figure out if your AI startup actually adds delta to the existing solutions.
1. Do You Have Proprietary Data?
AI is only as good as the data it learns from. OpenAI’s models are trained on terabytes of public data. Most AI wrappers don’t stand a chance because they rely on the same public datasets everyone else is using. But most of the world’s valuable data lives behind firewalls, locked up by enterprises or collected by specialized players.
You can create a strong competitive advantage by enabling small language models (SLMs) to serve highly specialized tasks more efficiently than commoditized, general-purpose models. This is especially true in industries with sensitive or niche data, like finance, healthcare or specific market insights.
Autonomous driving data is a textbook example. Cars from companies like Tesla generate large amounts of real-world driving data daily. Competitors can’t just scrape that off the internet.
If your startup has access to unique, valuable and scalable datasets, you’re already miles ahead.
2. Where Are You In The AI Stack?
Not all AI startups are created equal. Where your idea fits in the stack matters.
• Foundational Layer: Building the model itself (think OpenAI and Anthropic). These players typically win on scale and R&D muscle.
• Middleware Layer: Think LangChain or Hugging Face—these startups enable others to use, tune or deploy models efficiently.
• Application Layer: Here’s where most startups fall. If you’re building at this level, you need laser focus on a niche vertical, data access or an exceptional UX to stand out. Canva, for example, adds AI to a robust application layer for intuitive design.
For those in foundational or middleware layers, accuracy, reliability and latency are paramount.
3. How Will You Manage Costs?
AI is compute-heavy. Training large models burns money like a VC fund in a bull market. Can you offer cost advantages over competitors? Do you have a plan to make your costs manageable? Partnerships, clever infrastructure choices, or simply a scalable pricing model can be the difference between survival and an early exit.
4. Can You Leverage The Network Effect?
The more your solution integrates with existing tools, the harder it will be to replace. Look for opportunities to become embedded in workflows. This compounding environment makes it hard for competitors to catch up.
Salesforce, which started as a CRM, has morphed into a deeply integrated platform with thousands of third-party apps embedded into its customers’ workflows.
Before You Go All In…
Ponder these questions for a while:
• Does your idea solve a repeatable, real-world problem?
• Do you have a clear advantage over what’s already out there?
• Will your economics hold up when the VC money slows?
• What do you have that incumbents don’t?
These questions will help you identify if you’re building a product or a sustainable business.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here
