When Everyone Can Build, What Is Actually Defensible?

When Everyone Can Build, What Is Actually Defensible?

Imagine walking into an investor meeting with a working product.

You demonstrate it. It looks good. It solves a real problem. Something that might once have required a team of developers and months of work is sitting there, functioning, in front of them.

And someone on the other side of the table goes home and tries to rebuild it. Not because they want to steal your company, but because they want to know how difficult your company is to reproduce.

I recently heard an investor describe funds doing some version of this with AI-assisted development. A startup presents a demo; someone attempts to recreate its core functionality. If they can produce something similar relatively easily, that tells them something.

Or perhaps, more precisely, it tells them what the demo doesn’t prove.

Which raises a deceptively simple question: When everyone can build, what is actually defensible?

The artifact and the advantage may no longer be the same thing

We naturally look at what a company has made when trying to understand its value: the software, product, interface, feature, service.

• The artifact is visible. The capability behind it is harder to see.

If something was technically difficult to produce, the existence of the thing itself once provided evidence of unusual capability. But a beautiful interface might now be generated quickly. A sophisticated feature might be recreated. A functional prototype might be assembled in a weekend.

The artifact remains real. What changes is what we can infer from its existence.

Perhaps one increasingly important distinction will be between what a company has built and what made that company capable of building something worth having.

The first may be relatively easy to copy. The second may be considerably harder.

Yesterday’s difficulty may not remain tomorrow’s barrier

Companies have always benefited from barriers: technical complexity, capital requirements, specialized expertise, distribution, regulation, switching costs, proprietary intellectual property, network effects and customer relationships.

AI alters some of those equations. If software becomes faster and cheaper to produce, technical implementation alone may offer less protection. If small teams can suddenly access capabilities that once required large specialist teams, organizational scale itself becomes less protective.

None of this means defensibility disappears. It means yesterday’s difficulty may not remain tomorrow’s barrier.

Perhaps the moat is upstream of the product

Suppose two companies can build roughly equivalent technical products.

One understands the customer considerably better. One has spent years inside the industry. One has relationships the other doesn’t. One possesses proprietary data. One has discovered a problem that customers themselves struggle to articulate.

One learns faster.

Those advantages may eventually appear in the product, but they do not originate there.

Perhaps the moat is upstream of the product.

• If implementation becomes easier, deep understanding of the system you’re intervening in may become more valuable. Someone who knows exactly where the problem lives can use increasingly accessible technical capability to attack it.

The technology matters. But the advantage may live in what directs it.

Data can be copied. What have you learned from it?

Proprietary data frequently appears in discussions about AI defensibility, for obvious reasons. But “we have data” can become its own comforting answer.

What kind of data? How difficult is it to acquire? Does it improve through use? Can competitors obtain something equivalent? Does it actually produce better decisions?

And most importantly: What has the organization learned from it? A company that has spent years observing a particular problem may know which signals matter, which apparent correlations mislead, which customer requests should be ignored, which approaches repeatedly fail and which exceptions reveal something important.

That isn’t merely proprietary data. It is proprietary learning.

• The more interesting moat may therefore be not simply owning information competitors lack, but developing a way of seeing the problem that competitors haven’t yet developed.

Some of the slowest things to build may become the hardest to copy

There is another form of advantage that looks almost embarrassingly old-fashioned in a conversation about artificial intelligence.

Relationships.

Trust still takes time. Reputation accumulates through repeated experience. A customer who believes you understand their world is not automatically transferable to the next company that reproduces your feature. A community cannot always be generated on demand. A network is not simply a database of names.

The faster technology allows companies to move, the more noticeable these slower forms of advantage may become.

• Some of the most defensible assets in an accelerated company may be the ones that cannot be accelerated very much at all.

Speed matters. But what are you learning while moving quickly?

If one company can test ten ideas while another tests two, the first has more opportunities to learn.

• But AI is giving everyone more speed. Eventually speed itself may become less differentiating.

The better question may be: What are you learning while moving quickly?

A company that can observe, interpret, decide, test and learn faster than competitors possesses something more interesting than production velocity: it has learning velocity.

And that may prove harder to copy because it isn’t located in a single tool. It lives across the organization’s habits, relationships, information and decisions.

Brand may matter more when products look more alike

Brand is sometimes treated as the decorative layer applied after the important work is done: logo, colours, voice, campaign. But if functional differences between products become easier to reproduce, meaning may carry more weight.

Why this company? Why this approach? Why should I believe you? Why would I choose you when another company can demonstrate similar capability?

A brand cannot rescue a product nobody wants. But neither is brand merely the visual surface of a product.

At its strongest, a brand is an accumulation of expectations.

• A competitor can reproduce your feature. It cannot instantly reproduce what people have come to believe about you.

Some advantages begin with an idea other people don’t believe yet

A company may also possess an unusual understanding of the world.

A founder notices something others have missed. Customers behave in a way conventional wisdom says they shouldn’t. A technology makes possible something that recently wasn't.

That kind of insight can create advantage precisely because it has not yet become obvious. But if it succeeds, competitors eventually learn.

Today's unusual insight becomes tomorrow's conventional wisdom. Which points toward a larger problem:

What if defensibility isn’t something you build once?

Data can become available elsewhere. Employees leave. Patents expire. Competitors learn. Relationships weaken. Customer expectations change. Technology eliminates barriers.

Perhaps the search for a permanently defensible advantage is itself misguided. Maybe defensibility is increasingly dynamic.

The question changes from: What do we possess that nobody can copy? to

• What allows us to keep creating advantages faster than others can reproduce them?

That shifts attention away from protecting a static artifact and toward cultivating the conditions that repeatedly produce valuable differences: deep understanding, strong relationships, distinctive knowledge, learning, judgment, experimentation and the ability to notice what is changing before the answer becomes obvious.

So what did you actually build?

Which brings me back to the investor rebuilding the startup demo. The revealing question is not simply: Can we copy your product?

Almost everything can eventually be copied.

It is: If we copy what you’ve made, have we copied what makes you valuable?

If the answer is yes, the company may have a problem. If the answer is no, then the interesting work begins.

Perhaps the advantage lives in data, relationships, hard IP, accumulated learning, trust or an unusual understanding of the market. Or perhaps the most defensible thing a company can build isn’t a thing at all.

It is the capacity to keep seeing, learning and creating what others cannot easily reproduce.

::::::::::::::::::::::::::::::::::::::::

Contemplations & Provocations | A field essay by
Marie Fe Isla Rae

© 2026 Marie Fe del Rosario. All rights reserved.
Terms of Use · Privacy Policy · Permissions & Collaborations

All content on this site—including text, frameworks, methodologies, visuals, and strategic concepts—is protected by copyright and may not be reproduced, adapted, or used without prior written permission.

Authorship and web design by Marie Fe.

Field notes on consequential moments, creative strategy, experience, narrative, design—and the work of making possible futures tangible.

For people interested in what happens between imagining and making.