When AI makes the work cheap, ownership matters
AI agents are making knowledge work cheaper, faster, and more abundant. This isn't a doomsaying prophecy, but a fundamental shift Greg Isenberg identifies as the "agentic era," where the advantage moves decisively to those who own assets rather than simply selling hours.
This thesis isn't a guarantee of universal prosperity. Abundance doesn't inherently mean profit for everyone. Durable value still hinges on customer demand, flawless execution, and defensible advantages in a hyper-efficient market. But the rules of engagement are changing.
The crucial question becomes: how do you pair AI's unprecedented leverage with tangible assets? This piece explores business models that integrate AI with:
- Trust
- Audience
- Data
- Specialized workflows
- Physical experiences
Think of Othership, the Toronto-born sauna club scaling in New York, Manhattan. AI might handle their admin, but the core product is an irreplaceable, in-person communal experience. Their asset isn't just the physical space, but the curated atmosphere AI can’t replicate. We'll examine these resilient archetypes, revealing where true value lies when AI makes the labor nearly free.
Sell the finished job, not another AI seat
AI-native service firms redefine value, shifting from billing hours to delivering completed outcomes. Greg Isenberg notes these firms leverage AI agents for 90% of routine delivery, with humans reviewing exceptions and ensuring quality. This model allows lean teams to serve more clients with unprecedented efficiency.
Consider niche bookkeeping: agents prepare client books, humans review anomalies, and the firm charges for the completed monthly books—not for hours spent or software access. Similarly, creative services can offer "logos as a service" or "ads as a service," delivering defined deliverables at a fixed price rather than an hourly rate. This outcome-led approach generates high margins, often mirroring software’s 70-90% gross profit.
The opportunity is immense for specialized, reliable operations. However, this model also presents a constraint: generic agencies are easily copied. Success hinges on deep expertise, robust operational consistency, and cultivated customer trust. Selling the finished job, rather than another AI seat, ensures firms capture value from the abundance AI creates, proving that ownership of the deliverable—not just the tools—is paramount.
The internet is crowded; attention is an asset
Generating products with AI will soon be cheap, but earning attention and customer trust will become the next bottleneck—and more expensive than ever. Greg Isenberg argues that as AI makes creation abundant, the ability to distribute and capture an audience becomes the most valuable asset.
This shift prioritizes the Audience, Community, Product (ACP) funnel. First, publish useful niche information to gather an engaged audience. Next, foster that audience into a community. Finally, build products addressing its specific needs. This model flips traditional product development on its head, ensuring demand before supply.
Consider the cost of reaching customers: a Facebook ad click that once cost 5 cents in 2009 now averages $1–$5, depending on targeting. Owning organic audiences becomes a moat against rising ad costs and the noise of AI-generated content.
Furthermore, proprietary datasets and domain-specific harnesses offer a distinct advantage. While generic models proliferate, unique, high-quality information and industry-specific workflows can make an AI service far more useful than a generalized counterpart. Access rights and ongoing data quality are essential. For more on the rise of specialized services, see Services: The New Software. This approach capitalizes on the fact that not all data is created equal, making specialized information an asset AI can't easily replicate or commoditize.
Enjoying this? Get one like it in your inbox each morning.
one email a day · unsubscribe in two clicks · no third-party tracking
Bet on what a chatbot can’t deliver
Bet on what a chatbot can’t deliver: the physical, the communal, the truly human. As digital services overflow, the anti-AI wedge is the IRL experience. Othership, a Toronto-born communal sauna, exemplifies this, offering wellness, atmosphere, and membership in a shared physical space that no algorithm can replicate.
Such ventures, while compelling, demand significant capital, especially in high-rent markets like New York or Manhattan. Isenberg wisely advises a pragmatic approach: test your concept as an MVP before committing. Othership’s founder began in his backyard, proving the model before scaling to costly real estate. This iterative, low-overhead start mitigates risk in a capital-intensive sector.
Beyond these experiential havens, other anti-AI opportunities exist. Isenberg briefly highlights robotics and physical AI, specialized physical products with devoted fans, and the underlying infrastructure of compute or energy. While each presents unique challenges, they represent tangible assets insulated from the digital deluge and its inherent commoditization.
Frequently Asked Questions
What is an AI-native service firm?
It sells a finished service, with AI agents doing much of the work and people handling quality checks and exceptions.
Why might offline businesses thrive in the AI age?
In-person experiences, community, and physical spaces offer value that AI cannot deliver through a screen.
What does audience, community, product mean?
It is a sequence: build an audience, bring its members together, then create products based on what they need.
Do you need to build AI models to benefit from this shift?
No. Opportunities may include services, distribution, data, sales, or niche physical products—not only model development.