One of the first things I look for when evaluating a startup is deceptively simple:
What is the technology actually doing for the customer?
Not what it could do.
Not how technically impressive it is.
Not whether the founders have built something that would make an engineer excited.
What does it actually change for the person paying for it?
This matters because I see a recurring pattern in startups, particularly AI startups.
The technology is impressive.
The architecture is sophisticated.
The founders have built something genuinely difficult.
And yet, when you ask what the customer gets from it, the answer becomes surprisingly fuzzy.
That’s a red flag.
Technology is not automatically product advantage
A difficult technical problem can be a great business opportunity.
But “difficult to build” and “valuable to customers” are two completely different dimensions.
You can spend 18 months building something technically extraordinary and still end up with a product that nobody particularly needs.
Conversely, some of the strongest products are built on technology that is relatively ordinary.
The advantage comes from understanding the customer, integrating into their workflow, reducing friction, creating distribution, or simply solving the problem better.
So when someone tells me about their technology, I mentally translate it into a different question:
“Why will the customer care?”
If the answer is still mostly about the technology, I start digging.
The questions I ask
When evaluating a startup, I tend to work backwards from the customer.
1. What was the customer doing before this existed?
If the answer is “nothing”, I want to understand why.
If the answer is “using Excel, WhatsApp and a person”, that’s often much more interesting.
2. What changes for the customer?
Does it save time?
Make money?
Reduce risk?
Increase conversion?
Remove a painful manual process?
Enable something that was previously impossible?
“Uses AI” isn’t an answer.
3. Why does this need to exist now?
Sometimes the technology genuinely creates a new possibility.
That’s interesting.
But sometimes the technology is simply being attached to an existing problem because it’s fashionable.
That’s different.
4. What happens if a competitor gets the same technology tomorrow?
This is probably my favourite question.
If the answer is “then we’re basically the same”, the technology isn’t your moat.
It may be an enabler.
That’s fine.
But then the moat needs to come from somewhere else.
Distribution.
Data.
Workflow.
Network effects.
Brand.
Switching costs.
Customer relationships.
Execution.
Something.
The technology can still be extraordinary
This isn’t an argument against deep technology.
Some companies absolutely should be building difficult technology.
If you’re working at the frontier of AI, semiconductors, robotics or another genuinely technical domain, the technology is the product.
But that’s a very different proposition from taking a powerful new technology and asking:
“What business can I build around this?”
In the latter case, I want to see the customer problem before I get excited about the technology.
Because I’ve learned to distinguish between two kinds of impressive:
“That’s really hard to build.”
and
“That’s really valuable.”
The best startups manage to make both statements true.
But if I have to choose, I’ll take valuable first.