Investor Preferences in AI Companies: What’s In and What’s Out

Recent years have seen a surge in investments in AI companies, with the technology dominating the landscape in Silicon Valley and beyond. However, not all AI companies are receiving the same level of attention from investors.

While it may seem like every company is jumping on the AI bandwagon these days, some startup concepts are falling out of favor with investors. VCs have shared insights with JS on what they are no longer interested in when it comes to AI software-as-a-service startups.

Current Trends in SaaS Investments

According to Aaron Holiday, a managing partner at 645 Ventures, popular SaaS categories for investors now include startups focusing on AI-native infrastructure, vertical SaaS with proprietary data, systems of action, and platforms deeply integrated into critical workflows.

However, Holiday also highlighted the types of companies that are considered unappealing to investors at present. These include startups working on thin workflow layers, generic horizontal tools, light product management, and surface-level analytics — essentially, anything that can be easily handled by an AI agent.

Abdul Abdirahman of F Prime noted that generic vertical software lacking proprietary data moats is no longer attractive to investors. Igor Ryabenky, founder and managing partner at AltaIR Capital, emphasized the importance of product depth, stating that investors are seeking companies with substantial differentiation beyond just UI and automation.

Key Considerations for AI Startups

Ryabenky emphasized the need for new companies to focus on real workflow ownership and a deep understanding of the problem they aim to solve. He highlighted the importance of speed, focus, adaptability, and flexible pricing models in the current market environment.

Jake Saper from Emergence Capital discussed the significance of workflow ownership, pointing out the difference between products like Cursor and Claude Code. Saper noted that developers are increasingly prioritizing execution over process, which has implications for products aiming to attract human users in a world where AI agents are taking over workflows.

Integrations are also undergoing a shift, with Saper mentioning Anthropic’s model context protocol (MCP) as a game-changer in connecting AI models to external data and systems. The importance of being a connector in the market is diminishing, as integrations are becoming more streamlined and accessible.

Adapting to Changing Trends

As the landscape evolves, companies offering workflow automation and task management tools may find themselves facing challenges as AI-native startups with more efficient technology emerge. Ryabenky highlighted the risks associated with easily replicable SaaS products and emphasized the need for depth, expertise, and integration of AI in products.

Investors are increasingly favoring businesses that own workflows, data, and domain expertise, while steering away from products that can be easily replicated. Ryabenky advised companies to focus on integrating AI deeply into their products and updating their marketing strategies accordingly.

Overall, the key to attracting investor interest in the current climate lies in offering depth, expertise, and innovation in AI technologies within critical workflows. Companies that can demonstrate these qualities are likely to stand out in a competitive market.