A founding principle of our Enduring GROWTH System™ Research Program is that we’re learning from people who have built growth in very different contexts, including very different technology eras.
This week, we sat down with Somen Mondal, a serial founder with three exits behind him who is now building again as an AI-native company. We wanted to explore what changes, and what doesn’t, when the tools for building a company change this dramatically. Our conversation led to an intriguing question:
AI is changing how companies are built. But is it changing what makes them successful?
Somen is a serial technology founder who has co-founded and successfully exited three companies: Field ID, acquired by Master Lock; Ideal, an early AI company acquired by Dayforce (then Ceridian) in 2021; and FairNow, acquired by Optro (backed by Hg Capital) in October 2025. His companies span multiple technology eras, from SaaS and mobile through machine learning to today’s AI-native companies. Along the way, he has created new categories, scaled B2B enterprise businesses and navigated strategic acquisitions. He has done it without traditional venture capital, choosing capital-efficient growth funded by a trusted group of angel investors.
Today, as Co-Founder of Amp, he’s building differently. Somen draws a sharp line between his past and present companies. Ideal was an AI product, but AI wasn’t used to build or run the company. At Amp, AI touches nearly every process: building the product, marketing it and, eventually, selling it.
As Somen describes it, the difference is dramatic. Work that once took much larger teams can now be done by a handful of people. Product development is faster. Customer intelligence can be pulled together almost instantly. Marketing work that used to take days or weeks now takes hours. And in his experience, the relationship between capital, headcount and growth is starting to change.
Last week, historian-turned-growth-advisor Miguel Trujillo urged founders to “study what has been permanent,” and to tell the difference between what is changing the fundamentals of business and what is simply new. Somen gave us a chance to test that idea from the inside, through someone building through the change in real time.
Some of his most compelling observations weren’t about what AI changes. They were about what it doesn’t.
1. Customer proof still matters.
For all the ways AI is speeding up company building, Somen was unequivocal about what matters most at Amp’s current stage. In his words, customer proof is “the only thing that matters.”
Amp is working with design partners. These are customers who have the problem Amp solves and are helping shape the solution around their real-world needs. Somen’s first two companies each had essentially one key design partner. Amp already has seven or eight.
But getting customers to use the product isn’t enough. The goal is to turn those early relationships into measurable evidence that the solution works. Right now, that means building case studies Amp can take to market. Ideally, they come from recognizable brands that give the next prospect confidence that this new company, and this new category, is credible.
Somen learned this at Ideal, where the Canadian retailer Indigo became an important early proof point. At Amp, the team is more deliberate. They agree on logo use, case-study participation and reference calls with design partners up front, rather than asking after implementation.
Somen sees the central risk in the business today this way:
“How quickly can we turn our design partners into referenceable case studies?”
When AI makes it easier to build products and bring them to market, credible customer evidence becomes more valuable, not less.
2. Distribution still creates leverage.
Somen shared a line that sums up one of the biggest lessons from building several companies:
“First-time founders focus on the product. Second-time founders focus on distribution.”
Ideal made that lesson concrete. Large technology companies can offer more than referrals or marketplace exposure. With the right structure, a startup can draw on another company’s installed customer base, brand and sales organization.
Somen’s test for whether a partnership has become a real growth multiplier is practical:
“Are their salespeople paid to sell your product?”
If selling your product helps a partner’s salesperson hit quota, the relationship has gone well beyond an ordinary partnership. Their commercial engine is now helping drive yours.
Ideal reached that point with Dayforce, which white-labelled its technology and later acquired the company. SAP also became an important channel. Somen sees white-label as potentially even more powerful than resale. Your capability gets packaged into the larger company’s offering and can show up in every relevant deal.
Those relationships didn’t replace proving the business first. Ideal was about five years in, had been through a pivot and had substantial enterprise customers before the Dayforce relationship became possible. Somen notes that in an AI-native world, that timeline may shrink dramatically.
The technology may be changing. The leverage available through distribution isn’t.
3. Capital works best once you’ve figured out the machine.
Somen has built his companies differently from many technology founders. He has never raised institutional venture capital. Instead, he has raised fairly modest amounts from angel investors he has come to trust across several companies.
He doesn’t think outside capital is bad. He thinks its timing and purpose matter enormously.
His approach is to raise enough to get through the riskiest period and reach the next meaningful stage. Along the way, he is deliberate about creating good outcomes for founders, employees and investors.
The turning point comes when the company has figured out how its growth works. Once there’s a clear sales opportunity and a repeatable formula, more capital becomes extremely valuable, because it speeds up something the company already knows how to do.
His principle is simple:
Don’t use other people’s money to figure out the machine. Once you’ve figured out the machine, use other people’s money to fuel it.
He also warns about the opposite trap. Too much capital can make pivoting a luxury. It lets a company keep searching indefinitely instead of forcing it to prove it has found something customers value.
This principle keeps surfacing in our research. Jeff Wald described every raise as a window, with a few things that must become true before it closes. Miguel Trujillo described the shift from investment leading to proof to proof leading to investment. Somen arrives at the same place from the opposite direction: he has never taken institutional venture capital at all.
The principle may matter even more for AI-native companies. If companies can build products, operate and learn with much smaller teams, the capital needed to reach meaningful proof may fall too.
So the question for founders becomes less “How much can we raise?” and more “What have we proven, and what would more capital let us speed up?”
4. When almost anyone can build, knowing what to build matters more.
This may be one of AI’s more counterintuitive effects: the easier technology makes building, the more valuable domain expertise becomes.
Somen sees that firsthand at Amp. AI has dramatically changed how much each employee can produce. Product development is faster. Operations are more efficient. Customer conversations, advisor input, competitive research, product information and go-to-market data can increasingly be brought together and analyzed as a kind of company-wide “superbrain.”
The barrier to execution is falling.
But that doesn’t answer the harder questions:
- What problem is actually worth solving?
- What should you build?
- What are customers really telling you?
- Which signals matter, and which are noise?
Somen believes this makes domain expertise more important, not less. AI can increasingly help a company find and execute an answer. But it doesn’t eliminate the need for domain expertise and human judgment about whether you’re solving the right problem.
Somen has watched technology cycles come and go. Each one created huge excitement about what had suddenly become possible.
And one thing really hasn’t changed:
Someone still has to buy it.
What we’re taking from the conversation
Somen’s perspective sharpens an important distinction for our Enduring GROWTH System™ research.
AI is clearly changing how a company moves through the stages of growth: how fast, with how many people and on how much capital. What’s less clear — and increasingly interesting to us — is whether it changes what companies ultimately need to prove as they progress:
- Evidence that customers value what you’ve built.
- Leverage beyond the founders themselves.
- Enough proof to know when to accelerate.
- The judgment to decide what’s worth building in the first place.
The tools are new. The tests may be more enduring than we think.
A sincere thank you to Somen Mondal for sharing what he has learned building across several technology eras, and for giving us an early look at what may be different, and what may prove surprisingly enduring, as AI-native companies take shape. His insights are helping shape our understanding of what it takes to build companies that don’t just grow — but endure.
To learn more from Somen, connect with him on LinkedIn:
https://www.linkedin.com/in/somenmondal/
To learn more about Amp, visit their website here: https://runamp.com/