If you work in or around software right now, you’re feeling the pressure: AI headlines getting louder, not enough clarity on what to do about it, and a suspicion that the landscape is shifting faster than anyone can keep up with. Some of that pressure is warranted. Some of it is pointed at the wrong thing.
Yes, AI can now clone software in weeks instead of years. Over 40% of code written today is AI-generated, and enterprises are already replacing SaaS tools with custom-built alternatives. The technology layer is getting commoditized fast.
But your product’s value was never really in the code. It’s in the customer relationships you’ve built, the proprietary data running through the product, the workflows you’re embedded in, the partners and distribution you’ve earned, the trust that gets you invited into deals. None of that can be cloned.
The honest question isn’t “how do we keep up with AI?” It’s “do we actually know what’s hard to copy about our business, and are we investing in those layers?” AI levels the playing field on software. It doesn’t level the playing field on businesses.
The Fear Is Real, But It’s Pointed at the Wrong Layer
A developer can vibe-code a survey tool in an afternoon. Nobody is vibe-coding their way into HIPAA-compliant healthcare analytics with five years of audit history.
That’s the distinction that gets lost in the headlines. Yes, AI can produce working software fast. Where it struggles is everything that makes software into a business: data quality, security, compliance, integrations, user adoption, change management. Coding turns out to be the easiest step in a long chain of work. If the only thing holding your competitive position together was “we built some software,” you didn’t really have one.
But if your product is woven into how customers operate, if it holds data they can’t get anywhere else, if it’s earned trust over years of delivery, that position is far more durable than the current panic suggests.
The anxiety I hear in conversations with leadership teams is often pointed at the most visible layer: features, technology, product capabilities. That’s the layer AI is commoditizing. But it’s also the layer that was always the easiest to replicate, even before AI. What was hard to replicate before AI is still hard to replicate after AI. In many cases, it’s gotten harder, because AI can actually make those deeper advantages stronger if you know how to use it.
What’s Actually Hard to Copy
The same short list keeps showing up, in the research and in the deals I work on.
Customers, relationships, and go-to-market. You can copy an interface. You can’t copy five years of relationships across your install base. In real B2B buying cycles, decisions are shaped by trust: What do our peers use? Who has solved this for a company like ours? Whose team do we believe will show up when things go sideways?
I see this constantly in due diligence. When I’m assessing a SaaS platform for a PE buyer, the strength of the customer relationships tells me more about the durability of the business than any architecture diagram. High net revenue retention, low logo churn, multi-year contracts with expansion paths: those are signals the business has earned something an AI clone starts without.
And this extends beyond direct customer relationships to your partner ecosystem and the way you go to market. Your channel partners, your integration partners, your referral network, the warm introductions that get you into deals in the first place: those are hard to replicate and they compound over time. An AI clone has no distribution. It has no partners. It has no warm introductions.
Proprietary data and market insight. Every company likes to say “our data is our advantage.” In reality, only a small subset have data that’s genuinely proprietary, well-structured, and tied to outcomes. Those that do are playing a different game.
Foundation models are broadly available. The data you feed them is not. Ten years of customer behavior, transaction histories, and domain-specific edge cases will beat a technically elegant model with generic data almost every time. An AI-native startup can match your model architecture in a week. It can’t match your dataset in a decade.
Embedded workflows and switching cost. The more deeply your product becomes the nervous system of how a customer operates, the harder it is to rip out. Approvals, audits, handoffs, controls, integrations with adjacent systems: those are the real switching costs. A cloned product might look familiar in a demo, but replacing you means unwinding processes, retraining teams, rebuilding integrations, and accepting operational risk nobody wants to own.
I’ve managed product portfolios built through acquisition where some products were deeply embedded in customer operations and others were peripheral. The deeply embedded ones survived feature gaps, pricing pressure, and competitive noise because removing them was a business disruption, not a software swap. The peripheral ones were always vulnerable. That distinction matters more now than ever.
Brand, reputation, and community. For many buyers, choosing software is also a way to transfer risk. They’re buying your track record and your willingness to stand behind outcomes. A CFO choosing their first AI-powered FP&A tool isn’t running a feature comparison. They’re asking their CFO friends what they use and whether they trust it.
AI can generate content about trust. It can’t generate trust.
Domain expertise and scar tissue. Vertical and regulated markets run on accumulated learning from every edge case, failure, and regulatory change you’ve lived through. That expertise is embedded in product decisions, defaults, documentation, playbooks, and services. A general-purpose model can produce generic answers. It won’t recreate ten years of hard-won judgment in a specific niche.
Notice what all of these have in common. Every one of them was built by people. Your customers stay because of the humans who earned their trust. Your data exists because people made the decisions that generated it. Your workflows are embedded because someone sat with the customer and designed them. The product holds all of this together, but people created it and people keep it compounding.
And here’s what follows from that: those same people, the ones who know your customers, your data, your market, your operations, are also the ones best positioned to figure out how AI actually applies to your specific business. Better than any outside framework. Better than any AI-native startup that doesn’t know your customers. Your team carries the institutional knowledge required to make smart AI decisions. If you’re looking externally for someone to tell you where AI fits before tapping the people who know the business best, you might be overlooking your biggest advantage.
We’ve Been Here Before, But It’s Moving Faster
Nobody has a crystal ball on AI. This is new territory for everyone. No one has been through this specific wave before, and anyone claiming they have a proven playbook is selling you something.
But technology waves that reshape businesses aren’t new. When the internet arrived, the fear was that it would destroy traditional businesses. And it destroyed some. But the companies that thrived weren’t the ones who “got online” the fastest. They were the ones who understood what the internet actually changed about their customer relationships, their distribution, their market access, and evolved their business around it. The ones who treated it as a technology project (“we need a website”) mostly wasted money. The ones who treated it as a business evolution won.
The same pattern played out with agile and cloud computing. The technology was real, the disruption was real, and the winners were the ones who understood what was actually changing about their business, not just their technology.
AI follows that pattern, but compressed. The rate of change is faster, which makes it feel more urgent. And that urgency creates pressure to “do an AI transformation,” ship some initiatives, and check the box.
That instinct is the wrong one. And this is the part I want to be direct about.
AI is not a transformation with a finish line. It’s a continuous evolution. The technology will keep changing. The competitive dynamics will keep shifting. What AI can do today is not what it will do in eighteen months. If you treat this as a project — something you execute and then move on from — you’ll find yourself having the same panicked conversation again in two years, except further behind.
I’ve led organizations through enough transformational and evolutionary changes to know that the ones who treat a shift as a one-time transformation almost always end up in the same place: they did the work, checked the box, went back to normal, and got surprised when the world kept moving. The ones that internalized the change as a continuous way of operating pulled ahead and stayed ahead.
There is no “done” with AI. There’s only whether your organization is evolving faster or slower than the landscape around it.
The Honest Question
Here’s what I’d ask if I were sitting across the table from you right now.
Do you actually know what’s hard to copy about your business? Not what you’d put on a slide, but what you’d bet on if a well-funded competitor showed up tomorrow with an AI-built version of your product and aggressive pricing.
If the answer is “our features” or “our technology,” you have a problem, and you had it before AI showed up. If the answer is “our customer relationships, our data, the workflows we’re embedded in, our partners and distribution, the trust we’ve built, the people who know this market,” then you’re in a stronger position than the headlines suggest.
The discipline this moment demands is uncomfortable but straightforward: be honest about which layers of your product are genuinely hard to replicate, and which ones AI has already commoditized. Invest in the layers that matter. Build an organization and a team that can keep evolving as the landscape shifts. Because it will keep shifting.
AI levels the playing field on software. It doesn’t level the playing field on businesses.
In the next piece, I’ll get into what it actually looks like to use AI to strengthen your position rather than just checking a box.
Sources
- GitClear. AI Copilot Code Quality: 2025 Research. Analysis of 211 million lines of code from 2020 to 2024.
- Retool. The Build vs. Buy Shift: How Vibe Coding and Shadow IT Have Reshaped Enterprise Software (2026 report).
- Boston Consulting Group. AI Transformation Is a Workforce Transformation (2026).