For sophisticated AI tools requiring deep business context, a purely self-serve onboarding often fails. Plurium validates its PLG motion by initially using human consultants for setup to ensure data accuracy and gather context, then building those learnings into an automated, self-service flow over time.

Despite extensively using custom AI for interview analysis, Formation Bio finds that AI for candidate sourcing is still immature. Their talent team insists on a human reviewing every resume, highlighting that sourcing remains a significant automation challenge due to the need for nuance and confidence in evaluation.

For services like Secretary.com, the defensible moat isn't the AI model itself but the unique dataset generated by human oversight. This data captures the nuanced, intuitive reasoning of an expert (like an EA handling a complex schedule change), which is absent from public training data and difficult for competitors to replicate.

CEOs of platforms like ZocDoc and TaskRabbit are not worried about AI agent disruption. They believe the immense complexity of managing their real-world networks—like integrating with chaotic healthcare systems or vetting thousands of workers—is a defensible moat that pure software agents cannot easily replicate, giving them leverage over AI companies.

Instead of a generalist AI, LinkedIn built a suite of specialized internal agents for tasks like trust reviews, growth analysis, and user research. These agents are trained on LinkedIn's unique historical data and playbooks, providing critiques and insights impossible for external tools.

To ensure quality, The League's founder personally vetted every single applicant, rejecting those who didn't meet standards (e.g., 'gym selfies'). This unscalable, manual curation built immense trust and reliability, which became the brand's core differentiator and value proposition.

At massive scale, the product focus must shift from delight to trust. At LinkedIn, any change directly impacts users' economic opportunities, making risk mitigation the first principle. This contrasts with smaller products where prioritizing user delight and rapid innovation is more feasible.

To help families plan for complex long-term care costs, LTCareNav provides a tool allowing users to model the financial impact of various products, like insurance, on their own. By creating an educational sandbox free of sales pressure, the platform builds trust and empowers users to make informed decisions before they ever speak to a provider.

Dr. Fei-Fei Li asserts that trust in the AI age remains a fundamentally human responsibility that operates on individual, community, and societal levels. It's not a technical feature to be coded but a social norm to be established. Entrepreneurs must build products and companies where human agency is the source of trust from day one.