Early versions of AI-driven products often rely heavily on human intervention. The founder sold an AI solution, but in the beginning, his entire 15-person team manually processed videos behind the scenes, acting as the "AI" to deliver results to the first customer.
Related Insights
The popular belief that AI companies are inherently more efficient is a misinterpretation of their age. They are still hiring at a rapid rate. Human-intensive functions, like building a large enterprise sales force, still require significant time and headcount to scale, regardless of AI's influence on product development.
Artist's co-founder warns that the biggest mistake founders make is building technology too early. Her team validated their text-based learning concept by manually texting early users, confirming the core hypothesis and user engagement before committing significant engineering resources.
Despite hype, true 'autonomous marketing' is not imminent. AI excels at automating the first 80-90% of a workflow, but the final, most complex steps involving anomalies, nuance, and judgment still require a human. This 'last mile' problem ensures AI's role will be augmentation, not replacement.
To successfully automate complex workflows with AI, product teams must go beyond traditional discovery. A "forward-deployed PM" works on-site with customers, directly observing workflows and tweaking AI parameters like context windows and embeddings in real-time to achieve flawless automation.
Most companies are not Vanguard tech firms. Rather than pursuing speculative, high-failure-rate AI projects, small and medium-sized businesses will see a faster and more reliable ROI by using existing AI tools to automate tedious, routine internal processes.
Early enterprise customers won't invest time to become proficient with a complex data tool. Founders must join their meetings, operate the software for them, and surface insights to demonstrate value. This manual "data monkey" role is crucial for driving initial adoption.
Enterprises struggle to get value from AI due to a lack of iterative, data-science expertise. The winning model for AI companies isn't just selling APIs, but embedding "forward deployment" teams of engineers and scientists to co-create solutions, closing the gap between prototype and production value.
Off-the-shelf AI models can only go so far. The true bottleneck for enterprise adoption is "digitizing judgment"—capturing the unique, context-specific expertise of employees within that company. A document's meaning can change entirely from one company to another, requiring internal labeling.
Even powerful AI tools don't produce a final, polished product. This "last mile" problem creates an opportunity for humans who master AI tools and then refine, integrate, and complete the work. These "finisher" roles are indispensable as there is no single AI solution to rule them all.
Previously, building 'just a feature' was a flawed strategy. Now, an AI feature that replaces a human role (e.g., a receptionist) can command a high enough price to be a viable company wedge, even before it becomes a full product.
