What makes a business defensible when AI can build anything overnight

When AI can build anything overnight, the businesses that win are the ones with moats that can't be coded.

What used to take engineering teams months now takes days. Sometimes hours. AI has compressed the timeline for building software, launching products, and entering markets to a degree that would have seemed implausible even two years ago.

A recent podcast with Paul Bassat surfaced some points worth unpacking. Paul co-founded SEEK in 1997, then Square Peg, the venture capital firm behind early investments in Canva, Airwallex, and a string of other companies that have shaped the Australian tech landscape. He's spent three decades watching technology waves reshape industries. When he talks about what makes businesses defensible, the room tends to pay attention.

For finance leaders watching AI unfold, the implications cut both ways. New competitors can spin up overnight. But the advantages that actually count in this environment might be exactly what finance teams already have.

The question: what does defensibility look like when the playing field has fundamentally shifted?

When everyone can build fast, speed stops being the advantage

The barrier to building has collapsed. Tools that required dedicated teams and six-figure budgets now sit accessible to anyone with a laptop and an idea. A solo founder can ship in a weekend what a startup couldn't ship in a quarter five years ago.

This recalibrates the competitive equation in ways that aren't immediately obvious.

If anyone can build, building ceases to be the moat. Speed to market, once a defining advantage, becomes table stakes. The businesses that pull ahead will be the ones with advantages that can't be replicated by prompts and weekends.

That's the new strategic reality. Speed combined with structural advantages is what actually wins.

What investors look for has shifted in the past eighteen months. Execution used to be everything, and execution still counts, but strategic agility has become even more important. We're in a much more dynamic world where none of us know how the market will evolve, which means the people who get the big prizes are the ones who can adapt as they learn rather than locking into a fixed plan.

The moats that compound over time

If speed isn't the differentiator, what is? The answer comes down to assets and positions that accumulate value over time and can't be conjured from scratch.

The founders who build these moats share certain characteristics. They're obsessed with a problem, often doing what Paul describes as their life's work. They combine huge self-belief with enough humility to surround themselves with people smarter than them. And they have an insane level of energy sustained over years, because moats don't get built in sprints.

Data that sharpens the longer you operate

You can't vibe code your way to ten years of transaction history. Customer behaviour patterns, operational flows, payment routing decisions made millions of times over. All of this trains your systems on patterns that a day-old competitor simply cannot access. AI doesn't diminish the value of proprietary data, it amplifies it, which means the data moat actually widens the more AI matures. For example, A fintech processing US$150 billion in annual payment volume sits on a dataset that no prompt can replicate.

Regulatory and licensing architecture

Sixty-plus licences across global jurisdictions isn't something a startup replicates in a weekend. The labyrinthine complexity of financial regulation, often viewed as a burden, transforms into a structural barrier to entry. New competitors can build the software, but they cannot build the compliance infrastructure overnight, and regulators aren't going to accelerate their timelines because AI made the code easier to write.

Customer relationships forged through consistent delivery

In finance, where the cost of getting it wrong runs high, trust accumulates through years of reliable execution. Switching costs are real and relationships compound over time, which means a flashy new entrant with a sleeker interface still has to earn the trust that established players have built transaction by transaction. That takes years, not sprints.

Distribution at scale

Reaching customers remains genuinely hard. ChatGPT famously hit 100 million users in six weeks without spending on marketing, but that's the exception that proves the rule. For most businesses, distribution is a grind, and the companies that have already cracked it hold an advantage that AI doesn't erase.

None of these moats hinge on technology. They hinge on position, and position, unlike code, can't be copied overnight.

The question most companies are getting wrong

Most companies are asking the wrong question about AI.

"How do we use AI?" leads to tool-first thinking, where you adopt chatbots because competitors are adopting chatbots, and add AI features because AI features are expected. It's an activity that feels productive but doesn't address the underlying strategic question.

The sharper question is what problems can you now solve that you couldn't solve before?

Nobody asks "what should I do with my car?" or "what should I do with electricity?" These technologies became utilities we use to accomplish what we actually want, and AI will follow the same path. The real opportunity is identifying the problems that only became solvable once the technology arrived.

Building from strength

The AI era rewards clarity about what you actually have, not what you wish you had or what competitors seem to have, but what you actually have that's hard to replicate.

For founders, that clarity often points to advantages you've already built without realising their value. Years of customer data. Regulatory approvals that took eighteen months to secure. Relationships with suppliers and partners who trust you because you've delivered consistently. Distribution channels that took years to establish. These advantages predate AI and will outlast any particular wave of technology, and they form the bedrock of defensibility in an environment where everything else can be built overnight.

To navigate this environment you need conviction. You have to believe in something and stand for something, but endless debate teaches you nothing while action and observation teach you everything. If you're curious and thoughtful and you look at the data, you can tell whether your first hypothesis was right, partly right, or completely wrong, and then you iterate. The companies that thrive in the AI era will be the ones that understand their moats clearly and deploy AI to strengthen them.