The Untouchables: 12 Moats AI Will Never Breach

I’ve been going on about community being the answer to AI-driven commoditisation. I’m not wrong, but that’s not the entire picture.

Warren Buffett’s original idea of a “moat” referred to the structural advantages that protect a business from competitors. Today, AI is burning down many of the old moats, product features, content production, even brand aesthetics. And while community has emerged as a powerful new form of defensibility, it's not the only one. In fact, it isn’t the right moat for every business on its own.

Community-enabled defensibility is real. It’s durable. It’s human. But it works best when layered with other moats, just as Buffett intended.

AI will commoditise features. It will commoditise content. It will even commoditise expertise.

But it will not commoditise every structural advantage.

Here are twelve sources of defensibility that still matter, and in many cases, become even more valuable, in the era of AI.

Most teams underestimate this. They obsess over product quality while ignoring the harder truth: distribution beats product. Always has. Now it beats AI too.

Why AI can't commoditise it

AI can copy features, generate content, and replicate functionality. It cannot copy:

AI tools can't show up in people's inboxes unless someone chooses them. They can't appear in someone's feed unless they've earned attention. Distribution is control of attention. Attention is defensibility.

Practical forms

A newsletter with 50,000+ deeply aligned readers creates leverage no AI can replicate overnight. Strong SEO moats and ranking leadership compound over years. High-converting YouTube or TikTok channels become owned media assets. Partnerships with ecosystems become repeat pipelines. Paid acquisition engines optimised with proprietary intent signals outperform generic targeting.

The strategic reality: The best product doesn't win. The product with the best distribution wins. If you control the channels, you control the game.

Over the next year, Roku predicts that 100% of the streaming audience will see ads. For growth marketers in 2026, CTV will remain an important “safe space” as AI creates widespread disruption in the search and social channels. Plus, easier access to self-serve CTV ad buying tools and targeting options will lead to a surge in locally-targeted streaming campaigns.

Read our guide to find out why growth marketers should make sure CTV is part of their 2026 media mix.

Learn more.

This gets dismissed as "soft." That dismissal is a strategic error.

Brand is psychological defensibility. It operates at a level AI cannot touch.

Why AI can't commoditise it

AI can generate infinite alternatives. But consumers don't want infinite alternatives. They want familiarity. Trust. Identity alignment. Confidence.

Brand reduces cognitive load. In a world of overwhelming choice, people choose what feels safe. They choose what they already believe in. They choose the name that means something.

Practical forms

Category leadership; being the default name people think of. A clear point of view or doctrine that attracts believers. Founder-led narrative that creates emotional connection. Strong cultural positioning that signals tribal alignment. Symbols, rituals, and language that become identity markers.

The strategic reality: People buy who they believe, not who has more features. Brand is a moat because belief is harder to replicate than functionality. AI can generate a thousand alternatives. AI cannot generate trust.

AI augments expertise. It does not replace it, especially when that expertise is hard-earned, experiential, embedded in people, and tied to relationships.

Why AI can't commoditise it

Knowledge is cheap. Experience is not.

AI can access and synthesise information faster than any human. But pattern recognition built over decades in a specific domain? Judgment formed through thousands of real-world decisions? Relationships cultivated over years of trust-building? These remain human advantages.

Practical forms

Regulatory expertise that takes years to develop. Deep industry operators who understand nuances no training data captures. Specialists with decade-plus pattern recognition in complex domains. Strategic operators with rare insight into how things actually work. Talent networks competitors cannot access or recruit fast enough.

The strategic reality: Your people become the differentiating layer. Not because AI can't do what they do—but because the combination of expertise, judgment, and relationships creates outcomes AI alone cannot replicate.

Most companies claim their data is a moat. They're usually wrong.

The only defensible data is hard to collect and impossible to reproduce. If a competitor could theoretically gather the same data by launching a similar product, it's not a moat—it's a head start.

Why AI can't commoditise it

Training data determines model quality. Unique data creates exclusive leverage. If you have data nobody else can access, your AI applications—or the insights you derive—will outperform anything built on generic datasets.

What counts as real proprietary data

Behavioural data from a niche segment accumulated over years. Longitudinal performance data tracking outcomes over time. Operational data from physical processes competitors don't have. Industry-specific datasets that require privileged access to collect. Enriched user knowledge bases built through millions of interactions.

The strategic reality: You can't fake ten years of user behaviour logs. You can't copy telemetry collected from millions of real interactions. You can't shortcut the accumulation of unique, hard-won data. If you have it, protect it. If you don't, start building it.

If you control the ecosystem, you control the rules. Everyone else plays inside your environment.

Why AI can't commoditise it

Platforms scale faster than products. Ecosystems increase switching costs more than features ever could. When third parties build on your infrastructure, they create value that attracts users, which attracts more third parties. This flywheel is extraordinarily difficult to replicate.

Practical forms

API ecosystems where developers build on your platform. Integration hubs that become central to how other tools connect. Marketplace models where buyers and sellers depend on your infrastructure. Standards or protocols you define that the industry adopts. Interoperability that locks others into your orbit.

The strategic reality: Once others depend on your rails, you're defensible. The cost of leaving isn't just switching products, it's abandoning an entire ecosystem of integrations, partnerships, and dependencies. AI can replicate a feature. It cannot replicate a platform others have built their businesses on.

The most ignored moat: being embedded into daily operations.

This isn't about being a good tool. It's about becoming infrastructure that people can't remove without operational disruption.

Why AI can't commoditise it

AI can make alternatives cheaper. It cannot make switching painless.

If your product becomes part of the workflow, the process, the compliance requirements, the reporting cadence, the internal approvals, then replacing you requires operational overhaul. That's friction no competitor can eliminate with better features.

Practical forms

Integrations that become operational defaults. Automations tied to internal systems. Mandatory audit or reporting steps that run through your platform. HR, finance, or legal processes with locked dependencies. Data flows that would require months to re-architect.

The strategic reality: Workflow lock-in beats feature lock-in. The deeper you embed into operations, the higher the cost of removal. Become infrastructure, not just a tool.

Regulation is a moat if you're already compliant and others are not.

This advantage is often overlooked because compliance feels like overhead. But in regulated industries, being first through the approval process creates barriers that can last years.

Why AI can't commoditise it

Compliance is slow, expensive, and often political. Regulators move at human speed. Approvals require relationships, track records, and institutional trust that cannot be accelerated.

If you've already navigated the regulatory maze, competitors face real friction. They need the same certifications, the same approvals, the same relationships, and they're starting from zero.

Practical forms

Industry certifications that take years to earn. Regulatory approvals with lengthy review processes. Licenses that limit market participants. Completed audits that demonstrate compliance history. Data residency compliance in multiple jurisdictions. Government relationships and contracts that create incumbent advantage.

The strategic reality: AI cannot accelerate bureaucracy. Humans still run that domain. In healthcare, finance, defence, energy, and other regulated sectors, compliance isn't cost - it's moat.

In a world obsessed with software, physical assets often provide the most durable moats.

Why AI can't commoditise it

AI can simulate. It cannot replace real-world logistics, supply chain control, manufacturing capability, distribution networks, or specialised equipment.

Building physical infrastructure requires capital, time, permits, relationships, and operational expertise. There are no shortcuts. No prompt can conjure a warehouse network or manufacturing plant.

Practical forms

Cloud hosting infrastructure with global presence. Fulfilment centres enabling fast delivery. Hardware footprints deployed at scale. Robotics systems and automation equipment. Labs and R&D facilities with specialised capabilities. Distribution networks covering key markets.

The strategic reality: Software-only companies underestimate how powerful real-world infrastructure is as a barrier. Amazon's moat isn't the website, it's the fulfilment network. Tesla's moat isn't the software, it's the factories and charging network. Physical infrastructure takes years to build and billions to fund.

Not the unethical dark-pattern kind. This is about creating genuine value that users lose if they leave.

Why AI can't commoditise it

Switching isn't about effort. Switching is about loss.

AI can rebuild a feature. It cannot rebuild a decade of user-specific work. The history, configurations, integrations, and personalisation accumulated over years creates switching costs that no competitor can eliminate.

Types of defensible switching costs

Historical data users have accumulated. Custom configurations tailored to their needs. Automations and integrations connected to other systems. Personalised environments built over time. Long-term reports and benchmarks that lose meaning elsewhere. Community achievements and reputation that don't transfer.

The strategic reality: Make your product more valuable the longer someone uses it. Make leaving feel like losing years of accumulated investment. That's not manipulation - that's building genuine, compounding value.

Your position in the market is not a brand statement. It's a barrier to substitution.

Why AI can't commoditise it

If you occupy a category definition rather than a category slot, alternatives don't count as substitutes. You're not competing on features, you're competing on meaning.

Positioning is cultural, not functional. AI cannot change public perception at will. It cannot redefine how people think about a category. It cannot claim the mental territory you've already occupied.

Practical forms

Owning a category name, being synonymous with the solution. Creating a mental model others reference ("We are the X for Y"). Establishing a mission that the industry aligns behind. Rising to become the default choice that requires justification to reject.

The strategic reality: When people think of the problem, they should think of you. When they consider alternatives, those alternatives should feel like compromises. Positioning creates a frame that makes substitution feel wrong, not just different.

Money is not a moat. But capital deployed correctly is.

Why AI can't commoditise it

AI lowers the cost of building products. It does not lower the cost of building empires.

With sufficient capital, you can accelerate ecosystem development, infrastructure buildout, compliance processes, and distribution faster than new entrants can keep up. You can buy what others must build. You can outspend competitors into irrelevance.

Practical forms

Buying distribution through acquisition or exclusive partnerships. Buying exclusive data access others cannot afford. Acquiring niche capabilities before competitors recognise their value. Building physical or digital assets at scale others cannot match. Locking partnerships through economics competitors cannot offer.

The strategic reality: Capital is a multiplier, not a moat itself. But capital applied to other moats - distribution, infrastructure, ecosystem, compliance, creates compound advantages that become insurmountable.

Human relationships still matter. They always will.

Why AI can't commoditise it

AI cannot replicate trust built over years. It cannot replicate political relationships that open doors. It cannot replicate enterprise partnerships forged through shared history. It cannot replicate supplier loyalty earned through fair dealing. It cannot replicate human social capital.

In B2B, regulated sectors, and anywhere decisions involve significant stakes, relationships remain decisive. People do business with people they trust. That hasn't changed. It won't.

Practical forms

Long-term client relationships that survive competitive pressure. Regulatory relationships that smooth approval processes. Supplier partnerships that guarantee access and pricing. Industry relationships that create deal flow. Personal networks that open doors competitors cannot enter.

The strategic reality: Relationships are slow to build and impossible to shortcut. AI can help you manage relationships at scale. It cannot replace the trust, history, and human connection that makes those relationships valuable.

The most defensible businesses don't rely on a single moat. They layer multiple moats that reinforce each other.

The flywheel spins.

AI is coming for your features. It's coming for your content. It's coming for anything that can be reduced to patterns and replicated at scale.

It is not coming for your distribution, your brand, your relationships, your regulatory approvals, your physical infrastructure, or the ecosystem others have built their businesses on.

Build what cannot be copied. Layer what reinforces itself.

That's how you survive the AI wave.

Community-enabled defensibility is one powerful option. These twelve are others. Most businesses need a combination.

The question isn't which moat to build. It's which combination creates compounding defensibility for your specific business, and whether you're building it fast enough.

AI isn't waiting. Neither should you.

We appreciate you taking the time to read this. Building community isn't just theory for us, it's a framework we've developed and refined with companies at every stage.

Have questions about adopting The Moat framework for your business? We'd love to help. Book a discovery call to speak directly with one of our consultants. We'll audit your current community engagement process and show you exactly where the opportunities are.

Did someone forward this to you? Subscribe here to get strategic insights on building community as your competitive advantage, delivered directly to your inbox.

Keep building your community until next time.

The Moat
Where Connection Becomes Your Competitive Edge

This article was written with assistance from artificial intelligence (AI). For more information, please visit our AI Statement.