AI won't save a weak business model
AI is attracting most of the venture capital in the market. Adam Ryan argues that adding AI to a pitch does not fix a weak business model, and sets out the fundamentals he still looks for.
Key points
- Carta reported more than 60% of venture capital raised on its platform in Q1 2026 went to AI companies.
- The fundamentals of a good business, including pain, unit economics, retention and defensibility, still apply.
- AI can weaken a model through thin margins, weak defensibility and pilots that never convert.
- AI strengthens a model when it creates proprietary data, workflow ownership or operational leverage.
Artificial intelligence is the defining theme of today’s venture market. Carta reported that more than 60% of all venture capital raised by companies on its platform in the first quarter of 2026 went to AI companies. Founders feel the pull. Pitch decks that once mentioned AI on slide nine now lead with it.
I am enthusiastic about what AI can do. I use it, the companies I work with use it and some are built around it. But I want to make a simple point: AI does not fix a weak business model. If anything, it can expose one faster.
The fundamentals have not changed
When I assess a company, I ask the same questions I have always asked:
- Is there a painful problem that a specific customer will pay to solve?
- Can the company acquire customers at a cost that makes sense relative to what they pay over time?
- Do customers stay and spend more?
- Are gross margins healthy enough to fund growth?
- Is there something that makes the business hard to copy?
AI can help with any of these. It can make a product more useful, reduce the cost of serving customers or create data advantages. But it does not answer the questions on its own.
Where AI can make the model weaker
There are several ways AI can actually weaken a business model:
- Thin margins. If the product relies heavily on third-party models, the cost of each customer interaction can be high and unpredictable.
- Weak defensibility. If competitors can build the same feature using the same model, the advantage disappears quickly.
- Pilot purgatory. Customers are eager to experiment but slower to commit. McKinsey’s State of AI 2025 survey found that 88% of organisations use AI, but only about a third have begun scaling it across the enterprise.
- Pricing pressure. As AI capabilities become widely available, customers may expect lower prices.
Where AI genuinely strengthens a model
AI strengthens a business when it changes the economics in durable ways:
- Proprietary data that improves the product as more customers use it.
- Workflow ownership, where the product becomes the way work is done rather than an add-on.
- Operational leverage, where AI lets a small team serve many more customers.
- New products that were not possible before.
Questions I ask AI founders
When founders pitch an AI business, I ask what the product would look like without AI, how gross margins change as usage grows, what happens if a model provider raises prices or a competitor uses the same model, and which customers have moved from pilots to paid, scaled deployment.
The best AI founders I meet answer these questions directly. They talk about customers and economics first and technology second.
A view across markets
The AI investment wave is global, though concentrated in the United States. In Australia, Cut Through Venture found that companies building with AI attracted 61% of total funding in 2025. In Europe and Asia, AI is the leading theme for investors as well. Wherever a company is built, the same rule applies: technology is a tool, and the business still has to work.
Where I would focus
If I were building an AI company today, I would focus on a narrow, valuable workflow in an industry I understood deeply. I would invest early in the data, integrations and customer relationships that competitors cannot easily copy. I would track gross margins closely as usage grew and design pricing that reflects the value delivered rather than the cost of model calls. And I would measure success by customer retention and expansion, not by the number of pilots signed. Those habits matter more than which model sits underneath the product.
Frequently asked questions
Should every startup use AI?
Most will use AI in some way, but not every product needs to be built around it. Use it where it genuinely improves the customer outcome or the economics.
Are investors only funding AI companies?
AI takes a large share of capital, but investors still back non-AI companies with strong fundamentals. Competition for attention is intense.
What is the biggest risk for AI application startups?
Commonly, weak defensibility: if a competitor can replicate the product with the same model, pricing power and margins erode.
References
- Carta: State of Private Markets Q1 2026
- McKinsey: The state of AI in 2025
- Capital Brief: Aussie startups in 2025 had strongest funding quarter in years
Related reading
General information only. This article reflects the author’s views at the time of writing. It does not take into account your objectives, financial situation or needs, and it is not financial, legal or tax advice or a recommendation to invest. Figures are from the sources named and may change. See our terms of use.