AI Agents Require Better Management Systems, Experts Say
Venture capitalists and tech leaders say the next hurdle for AI agents isn't building smarter models-it's orchestrating them effectively in enterprise settings. Without proper management systems, companies risk fragmentation and control problems as they deploy multiple AI agents across workflows.
The challenge emerged in recent discussions between Aaron Levie, CEO of Box, and Steven Sinofsky, Board Partner at a16z. Both emphasized that individual agents performing isolated tasks deliver limited value. The real opportunity comes when agents collaborate, communicate, and delegate work to one another across an organization.
What Orchestration Actually Means
Orchestration goes beyond letting an AI respond to a prompt. It requires systems that can:
- Define clear workflows and task dependencies between agents
- Control which agents access sensitive data and systems
- Monitor agent performance and behavior for safety
- Enable agents to communicate and hand off work effectively
Sinofsky noted that current AI agent development often lacks these capabilities. Without them, organizations end up with disconnected tools rather than integrated systems.
The Shift From Simple Prompts to Complex Systems
Early AI work followed a simple pattern: send a prompt, get a response. Managing multiple agents interacting with different systems requires a different approach entirely.
Levie described the evolution as moving from asking an AI to perform a task to building systems where agents autonomously manage workflows. This demands understanding not just what the AI can do, but the broader business context and security implications.
Enterprise Adoption Faces Real Obstacles
Many companies are experimenting with AI agents, but integrating them into existing workflows presents significant hurdles. Data privacy, security, and governance models remain unresolved for most organizations.
The speakers highlighted that specialized tooling and platforms will be necessary to address these challenges. Companies that solve orchestration problems will gain competitive advantage.
APIs and Standards Matter
For agents to work together, they need standardized ways to communicate and exchange information. Well-defined APIs between different agents and existing software systems are essential.
The conversation pointed to a broader need: companies should think about AI strategy as interconnected systems requiring careful management, not isolated models.
Industry-wide standards for AI agent development and orchestration could help ensure interoperability, security, and reliability across platforms. Without them, organizations will struggle to deploy agents safely and effectively.
Learn more about AI Agents & Automation and AI for Management to understand how orchestration fits into your organization's strategy.
Your membership also unlocks:
