We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
MIT Technology Review·May 26·5 min read
Rethinking organizational design in the age of agentic AI
TL;DR
Organizations want to adopt agentic AI (autonomous software agents that execute workflows independently), but 76% lack operational readiness across people, processes, and infrastructure. The core problem is treating AI agents as add-ons to existing human-centered workflows rather than redesigning operations fundamentally.
✦ Why It Matters
Engineers must design AI agents for fundamentally redesigned workflows, not retrofitted legacy systems, to achieve intended performance and organizational value.
Key Takeaways
- Redesign workflows and organizational structure before deploying AI agents, not after.
- Audit current operations to identify which processes require fundamental reimagining versus incremental improvement.
- Involve people, process, and infrastructure teams simultaneously in planning agentic AI adoption.
- Define clear decision-making boundaries and autonomy levels for AI agents within new workflows.
Full Summary
Enterprise adoption of agentic AI—autonomous software agents capable of executing entire workflows with limited human oversight—is accelerating, with 85% of organizations targeting implementation within three years. However, a significant execution gap exists: 76% report their current operations and infrastructure cannot support this transition due to unpreparedness in people, processes, and workflows.
The fundamental issue, according to PwC's global CTO Prasun Shah, is that organizations are layering AI agents onto existing human-centered operating models rather than reimagining how work should be structured. This approach is compared to applying sticky tape to a breaking system—a temporary patch that prevents systemic improvement.
Agentic AI's actual value derives from agents' capacity to coordinate complex tasks, make independent decisions, adapt to changing conditions, and iteratively improve performance. Without organizational redesign at the systems level, companies risk disillusionment and fail to capture the transformative benefits these technologies offer.