Cisco Pushes AgenticOps to Fix Enterprise IT Fragmentation
The Development
Cisco is positioning AgenticOps as the solution to enterprise IT fragmentation, where operators currently navigate multiple dashboards and data silos to troubleshoot issues. DJ Sampath, Cisco's SVP of AI Software and Platform, warns this problem will intensify as employees deploy 10+ AI agents each. Cisco's approach centers on three principles: unified data access across network, security, and application silos; multiplayer-first design enabling cross-team collaboration; and purpose-built AI models trained on 40 years of operational data. The Cisco AI Canvas workspace replaces multiple dashboards with natural language interfaces, while the Deep Network Model provides domain-specific reasoning. Cisco is evolving its Duo platform beyond multi-factor authentication into comprehensive identity management for cross-domain agent access. The company emphasizes immediate action over waiting for AI markets to stabilize.
Our Take
Cisco is making a brilliant infrastructure play disguised as an IT operations story. While competitors focus on flashy consumer AI, Cisco is quietly positioning itself as the foundational layer for enterprise AI deployment—identity, data unification, and domain-specific models. The 40-year operational data advantage is real and defensible. More importantly, they're solving the actual enterprise AI problem: not getting better answers from ChatGPT, but managing the chaos when every employee has multiple AI agents accessing corporate systems. This positions Cisco to capture value from the enterprise AI boom regardless of which models or applications win. The identity layer evolution is particularly smart—whoever controls enterprise AI identity controls enterprise AI adoption.
The Exploit
Action Item
If you're planning enterprise AI agent deployments for 2026, evaluate Cisco's AgenticOps stack now before Q2 budget cycles close, as their unified identity and data access capabilities could eliminate months of custom integration work.