Agentic AI: From Assistants to Autonomous Teams

TLDR

• At EQT’s CEO Summit, Lovable, Bonterra, and Strawberry showed agentic AI moving from chat tools to systems that plan, act, and run workflows end-to-end.

Key Takeaways

• Lovable lets a product manager prototype full applications in minutes, compressing work that previously required engineering cycles.
• Bonterra framed agents as force multipliers rather than replacements, achieving efficiency gains without reducing headcount.
• Strawberry positions the browser as the interface layer for agentic work, already running end-to-end in sales, recruiting, and operations.
• Alexander Fred-Ojala, Head of AI at EQT Ventures, called 2025 the year of agents, citing systems now reliable enough to run coding tasks autonomously for hours.
• The implication: development, marketing, customer support, and finance can be restructured around outcomes delivered by swarms of AI agents.

Why It Matters

• The adoption framing matters as much as the technology: Bonterra’s internal rollout succeeded by positioning agents as amplifiers, not headcount cuts.
• Strawberry’s browser-native approach puts agentic capability on top of existing web workflows, requiring no back-end integration from the user.
• The shift from assistant to autonomous team changes what companies need to manage: human roles increasingly become orchestration of AI co-workers, not direct task execution.

· 2026-04-27 · Read the original