AI, software or humans

AI, Software, or Humans: Redefining Business Building in the Age of Automation

At Web Summit Vancouver 2026, Mr. Bryan Wish moderated a discussion on balancing AI, software, and human roles. Leaders must strategically allocate tasks, moving beyond AI simply replacing functions. Mr. Tim Springer, Co-Founder of Graviten, clarified AI enhances domain expertise, demanding expert human supervision. He views AI’s job impact as natural evolution. Ms. Deb Liu, Co-Founder of Ember AI, highlighted AI’s potential to free human time from routine tasks, allowing focus on unique human contributions like judgment and creativity. She stressed human wisdom in the loop, advocating for continuous feedback and auditing to ensure AI accuracy and prevent drift.

When integrating AI, Mr. Springer advised using human judgment for selecting AI-generated ideas, with human control points crucial for safety. Ms. Liu recommended involving frontline employees in co-creating AI solutions, leveraging their practical knowledge to design effective systems and build trust. This collaborative approach ensures automation addresses real needs and maintains organizational expertise.

Reflecting on his past, Mr. Springer, who built Level Access to $100 million, stated he would not rebuild it today. His criteria for a “good company” include solving a top-three pain for a budget-owning person, replacing existing spend, and ideally having the user and buyer be the same. Strong founder-market fit is also essential for long-term success.

Ms. Liu urged founders to prioritize understanding customer problems over solution development. Flexibility, adaptability, and persistence are vital. She advised failing fast and pivoting quickly if market demand is unproven, rather than investing years in a “default alive” company with minimal returns. This ensures resources are directed to validated opportunities.

Both speakers addressed scaling challenges. Mr. Springer noted that a “lovable code” prototype isn’t a full software company, especially for enterprise, which requires robust features like authentication and logging. Ms. Liu added that while AI lowers software development costs, the true competitive “moat” now lies in trust, network effects, reliability, and go-to-market strategy, not just the product itself. Extensive problem discovery and customer validation are paramount.

In conclusion, while the current ease of building software may lead to more average companies, there’s optimism for the future. Mr. Springer believes AI-mediated structured processes will accelerate the creation of better companies with solid foundations and strategic market entry. Ms. Liu reiterated that prioritizing customer understanding and feedback is essential for building truly great enterprises.