The Internet of AI Agents is the Future. In Conversation with Vijoy Pandey
Summary
Pulse of AI Podcast Host Jason Stoughton recently interviewed Vijoy Pandey, SVP and GM of Outshift by Cisco, Cisco's internal incubation engine, about the future of AI agents. Pandey, with a PhD in computer science and over 80 patents, discussed Outshift's vision for an "Internet of Agents" and the challenges in achieving seamless agent-to-agent collaboration. The conversation highlighted the need for open, interoperable platforms and protocols to enable agents to discover, identify, compose, communicate, and evaluate each other across various domains, including business, consumer, social, scientific, and physical work. Outshift has open-sourced much of its work in this area through the Agency project (AGNTcy.org), emphasizing the importance of tackling complex, "boring" problems for startups in the agent space.
Key takeaway
For CTOs and VPs of Engineering evaluating AI agent strategies, recognize that the future demands open, interoperable agent collaboration beyond enterprise silos. Prioritize adopting or contributing to open standards and protocols for agent identity and communication, such as those championed by Outshift's Agency project, to ensure your agentic solutions can scale and interact effectively across the broader ecosystem, mitigating future integration complexities and proprietary lock-in risks.
Key insights
The "Internet of Agents" vision aims for open, interoperable agent-to-agent collaboration across all human activities.
Principles
- Technology creates solutions and problems, driving a continuous innovation cycle.
- Open and interoperable ecosystems are crucial for broad participation and success.
- Identity should be a public commodity, not owned by a singular entity.
Method
Outshift's approach to agent collaboration involves four phases: discovery and identity, composition, communication, and evaluation/observability, supported by open-source protocols like A2A and MCP.
In practice
- Startups should focus on solving hard, "boring" business problems with agentic software.
- Leverage open-source projects like Agency (AGNTcy.org) for agent identity, discovery, and communication.
- Consider post-quantum cryptography for agent communication security from the outset.
Topics
- AI Agents
- Internet of Agents
- Outshift by Cisco
- Distributed Systems
- Open-Source Protocols
Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, AI Architect, Entrepreneur
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by The Pulse of AI.