GitHub and LF Networking Discuss Agentic AI Workflows for Software and Telecom

GitHub engineer Idan Gazit described “agentic workflows” that let AI agents execute software upgrades from simple English prompts, automatically parse changelogs, apply changes and open pull requests, boosting developer productivity. He emphasized that such automation requires robust safety guardrails to keep AI actions under control.

In a parallel view, LF Networking’s Ranny Haiby examined how the same agentic AI architecture could be applied to carrier‑grade telecom networks. He noted that both large‑scale software environments and networks share challenges such as distributed inventory, multi‑step workflows, telemetry analysis, policy compliance and the need for human‑in‑the‑loop escalation. While generic agent orchestration can be reused, the final “last mile” for networks demands domain‑specific engineering and strict guardrails, including protocols like the open‑source Agent2Agent (A2A) framework.