Lean teams running on AI will succeed or fail based on the humans directing the systems, argues Eric Francia, co-founder and CEO of Stockholm-based Uniplay, an agent-driven, game-based learning platform. With 59% of employees needing reskilling by 2030, according to the World Economic Forum, hiring for curiosity and autonomy matters more than hiring for a fixed skill set.
Why lean AI teams need questioning people
Francia, building his sixth company after scaling digital businesses across 70 countries, says knowledge made widely accessible by AI lets a handful of people complete in days work that once took a large team months. In that setting, companies become smaller and faster, and the choice of people trusted to guide AI systems becomes decisive. His prescription centers on workplaces where questioning orders is normal, not an exception tolerated from junior staff.
The approach rests on what he calls "AI by design" applied to people as well as technology. Trust in the human in the loop should match trust in the software, because the team shapes the company in ways founders cannot directly control. That means recruiting people able to make decisions when the founder is not in the room, and giving them real responsibility for strategic direction rather than abstract statements of vision.
The context is rapid change in job roles. A job description can differ from the advertised version by the time a candidate is found, and a six-month roadmap can be outdated once drawn up. Against that background, experience and academic credentials alone lose weight, while side projects and self-initiated builds become signals of curiosity. Curiosity, in this view, supports adaptation when technology overturns established processes.
What this means for companies using AI
For businesses adopting AI agents and automation, the practical shift is from skill-based hiring to learning-based hiring. Small firms gain speed when employees with autonomy spot product problems early and test fixes before issues take hold. Large firms gain engagement when staff participate in direction instead of executing narrow tasks. In both cases, responsibility from the start keeps company goals concrete for the people closest to the work.
The limits of this model deserve attention. Autonomy without clear ownership can blur accountability, and encouragement to challenge methods does not replace verification of results. Leaders still need to define decision rights, learning budgets, and review points for experiments. This account does not promise higher retention or faster delivery by itself; those outcomes depend on development investment and on whether new ideas receive structured follow-up.
A concrete marker to watch is whether reskilling plans for 2030 move from statements to programs with enrollment, completion rates, and role changes. Companies that publish such metrics and link them to smaller, AI-supported team structures will show whether autonomous humans deliver the expected speed. Absence of those disclosures will suggest the idea remains a hiring slogan.
