The AI gold rush is cooling for certain startups, and investors are getting picky! While billions are still flowing into artificial intelligence, not every AI company is a guaranteed ticket to investor riches anymore. Even with "AI" becoming the hottest buzzword, some startup ideas are now falling flat with venture capitalists. So, what exactly are investors not looking for in AI software-as-a-service (SaaS) companies these days?
The Hot List: What's Capturing Investor Interest Now
According to Aaron Holiday, a managing partner at 645 Ventures, the current darlings of the investment world include startups that are:
- Building AI-native infrastructure: The foundational tech that powers AI.
- Developing vertical SaaS with proprietary data: Software tailored for specific industries that has unique data advantages.
- Creating systems of action: Tools that actively help users get things done.
- Offering platforms deeply embedded in mission-critical workflows: Software that's essential for a company's core operations.
The Not-So-Hot List: What's Becoming a Snooze-Fest for Investors
Conversely, Holiday pointed out that investors are now finding these types of AI startups rather "boring":
- Thin workflow layers: Superficial tools that just add a bit of AI to existing processes.
- Generic horizontal tools: Software that tries to be everything to everyone, lacking specialization.
- Light product management: Tools that offer basic oversight without deep functionality.
- Surface-level analytics: Basic data reporting that doesn't offer significant insights.
Basically, anything an AI agent can now do with ease is a red flag.
The "Moat" Problem: Why Data and Depth Matter More Than Ever
Abdul Abdirahman, an investor at F Prime, highlighted that generic vertical software "without proprietary data moats" is losing its appeal. Igor Ryabenky, founder and managing partner at AltaIR Capital, elaborated, stating that investors are simply not interested in anything that lacks significant product depth. He emphasized, "If your differentiation lives mostly in UI [user interface] and automation, that’s no longer enough." The reason? The barrier to entry has dropped significantly, making it much harder to build a truly defensible business (a "moat").
The New Rules of the Game: Speed, Focus, and Flexibility
Ryabenky believes that new companies need to focus on "real workflow ownership and a clear understanding of the problem from day one." He added that massive codebases are no longer a badge of honor; instead, speed, focus, and adaptability are paramount. Pricing models are also shifting. Rigid per-seat models are becoming harder to defend, while consumption-based models are gaining traction.
Ownership is Everything: Workflow vs. Task Execution
Jake Saper, a general partner at Emergence Capital, offered a compelling analogy with Cursor and Claude Code. He explained that "one owns the developer’s workflow, the other just executes the task." He observed that developers are increasingly leaning towards tools that execute tasks rather than those that manage the entire process. This leads to a crucial question: If AI agents are increasingly handling the work, does "workflow stickiness" – the ability to keep human users engaged – still hold the same power as a competitive advantage? Saper mused, "Pre-Claude, getting humans to do their jobs inside your software was a powerful moat, but if agents are doing the work, who cares about human workflow?"
Integrations: From Moat to Utility?
Saper also pointed out that integrations are becoming less of a selling point, especially with advancements like Anthropic's model context protocol (MCP). This technology simplifies connecting AI models to external data and systems, reducing the need for multiple integrations or custom builds. "Being the connector used to be a moat," Saper remarked. "Soon, it’ll be a utility."
The Decline of Task Management and Generic Tools
Abdirahman echoed this sentiment, noting that workflow automation and task management tools that focus on coordinating human work are becoming less necessary as agents take over task execution. He pointed to public SaaS companies whose stock prices have dipped as new AI-native startups offer superior efficiency.
What Makes a SaaS Company Replicable (and Thus Unattractive)?
Ryabenky identified that SaaS companies struggling to raise capital are often those that are easily replicated. This includes:
- Generic productivity tools
- Project management software
- Basic CRM clones
- Thin AI wrappers built on top of existing APIs
He explained, "If the product is mostly an interface layer without deep integration, proprietary data, or embedded process knowledge, strong AI-native teams can rebuild it quickly. That is what makes investors cautious."
The Future of SaaS: Depth, Expertise, and Workflow Ownership
Ultimately, Ryabenky believes that what remains attractive about SaaS is depth and expertise, particularly in tools embedded within critical workflows. He advises companies to deeply integrate AI into their products and update their marketing to reflect this. "Investors are reallocating capital toward businesses that own workflows, data, and domain expertise," he concluded, "And away from products that can be copied without much effort."
But here's where it gets controversial... If AI agents are becoming so capable of executing tasks, does the focus on "workflow ownership" truly offer a sustainable advantage, or is it just a temporary reprieve before agents fully dominate? And for those building "generic productivity tools" or "basic CRM clones," is there still a niche for them if they can offer a truly exceptional user experience or a unique integration that AI agents can't replicate? What do you think? Let us know in the comments below!
