There Was Never An Easy AI Era, And Investors Are Raising The Bar

The belief that launching an AI product could secure quick funding has vanished. Small and medium-sized businesses enthusiastically embraced AI early on, but many struggled to retain customers after raising capital. These generic AI solutions, merely adding a layer of AI to existing software, often only provided marginal improvements that businesses were reluctant to adopt.

The AI market is now correcting this trend. Businesses that demonstrate tangible value are expanding faster than most enterprise software. Investors are now focused on which companies can maintain their customer base and whether forthcoming IPOs will reveal these differences. Defensibility in the AI business, according to Maor Farid, founder and CEO of Leo AI, hinges on three key factors.

First, the improvement must justify the investment. Saving a mere 10% on a sporadic task typically doesn't significantly alter a company's operations. Instead, reducing a critical business process from weeks to minutes, or saving substantial money, is what truly matters.

Second, domain expertise plays a crucial role. While advanced language models can now serve as infrastructure, the commercial potential lies in deeply understanding a specific industry to solve problems that generic AI tools can't. Third, proprietary context is essential. Companies building products with physical components often possess decades of accumulated knowledge, much of which remains hidden in outdated drawings or the minds of seasoned employees.

This industry-specific knowledge is absent from foundation models' training data. An AI product that lacks access to such context will struggle to become indispensable for the business.

As funding and IPOs for AI companies become more scrutinized, investors will closely examine metrics like net revenue retention and gross margins. They will assess if rapid growth translates into a sustainable business and whether each deployment becomes more cost-effective over time. Companies that can show expansion, meaning customers are increasing their spending post-implementation, will be seen as having proof that their product has truly altered how an organization operates.

Over the next 12-24 months, capital is expected to continue flowing to companies that combine robust industry expertise with substantial enhancements in critical business functions and access to proprietary knowledge. While some may succeed through broad self-service models, expanding within trusted customer accounts will likely prove more lucrative. Farid advises that growing steadily over five years, rather than experiencing a spectacular surge for just five quarters, is a wiser strategy.