Investors spill what they aren’t looking for anymore in AI SaaS companies

In recent years, billions have flowed into the realm of AI companies, reflecting the technology's immense influence across Silicon Valley and beyond. However, not every AI startup is capturing the attention of investors. As many companies rush to append 'AI' to their branding, several concepts are losing favor among venture capitalists. TechCrunch engaged with various venture capitalists to uncover the trends shaping investor preferences in AI software-as-a-service (SaaS) startups. According to Aaron Holiday, managing partner at 645 Ventures, the current hot categories for investors include AI-native infrastructure startups, vertical SaaS solutions with proprietary data, and systems designed to assist users in completing specific tasks. More critically, platforms integrated into essential business workflows are also drawing interest. Conversely, certain startup ideas have become less appealing. Investors are steering clear of those creating superficial workflow layers, generic horizontal tools, and basic product management solutions. In essence, any offering that can be easily replicated by an AI agent is now seen as less valuable. Abdul Abdirahman, an investor from F Prime, noted that generic vertical software without proprietary data is falling out of favor. Igor Ryabenky, founder and managing partner at AltaIR Capital, emphasized that products lacking substantial depth are unlikely to attract investors’ interest. "If your uniqueness is primarily in user interface and automation, that’s no longer sufficient," Ryabenky stated. He pointed out that with reduced barriers to entry, establishing a strong competitive edge has become increasingly challenging. New entrants must focus on owning workflows from the outset and thoroughly understanding the problems they aim to solve. He added that flexibility in pricing models is essential; rigid per-seat pricing will be harder to justify, while consumption-based models are gaining traction. Jake Saper, general partner at Emergence Capital, shared insights regarding workflow ownership. He described the distinction between products like Cursor and Claude Code as indicative of a broader trend. While one product manages the developer's workflow, the other merely executes tasks. Developers are increasingly gravitating towards solutions that prioritize task execution over process management. Saper warned that products focused on "workflow stickiness" may face challenges as AI agents become more prevalent in executing tasks. He noted that the historical advantage of engaging human users within a software environment may diminish as agents take over these functions. Furthermore, integrations are losing their allure, particularly as innovations like Anthropic’s model context protocol (MCP) simplify connections between AI models and external systems. Abdirahman highlighted that the necessity for workflow automation tools may decrease as agents autonomously handle tasks. He pointed to established SaaS companies whose stock values are declining due to the emergence of more efficient AI-native startups. Ryabenky remarked that SaaS firms currently struggling to secure funding are often those with easily replicable offerings, such as basic productivity tools or project management software. Ultimately, what's appealing in the SaaS landscape is depth and expertise, especially in tools that are embedded within critical workflows. Companies are encouraged to integrate AI deeply into their offerings and update their marketing strategies accordingly. Ryabenky concluded that investors are increasingly directing their capital toward businesses that possess ownership of workflows, data, and specialized knowledge, while moving away from easily replicable products.