{"id":1709,"date":"2026-07-29T09:41:26","date_gmt":"2026-07-29T14:41:26","guid":{"rendered":"https:\/\/portfolio.juandiegourdaneta.com\/blog\/sistema-operaciones-agente-ia\/"},"modified":"2026-10-20T20:57:52","modified_gmt":"2026-10-21T01:57:52","slug":"sistema-operaciones-agente-ia","status":"publish","type":"post","link":"https:\/\/juandiegourdaneta.com\/en\/proyecto\/sistema-operaciones-agente-ia\/","title":{"rendered":"AI Agent Operations System"},"content":{"rendered":"

A working system with an AI assistant for retrieving project information, organizing tasks and running automations.<\/p>\n

I configured Telegram access, persistent documentation and review and deployment workflows. The environment uses a Mac, private connections between machines and backups before website changes.<\/p>\n

I use it in my own projects. Decisions, permissions and results are recorded so work can be resumed and the assistant\u2019s actions reviewed.<\/p>\n

Automation supports operations; sensitive changes require supervision and verification.<\/p>\n

\n

OpenClaw, Telegram and operational tools<\/h2>\n

OpenClaw and existing agent tools integrate with Telegram as an interaction channel. Python and scheduled tasks support automation, while SSH and Tailscale provide managed connections between environments. The work consists of configuring and integrating these components, not claiming authorship of their frameworks.<\/p>\n

What this development contributes<\/h2>\n

Persistent documentation preserves decisions and resumption points. Separating reading, preparation and execution helps review the scope of each action and verify outcomes. For a business evaluating AI agents, this case shows why connecting tools with permissions and traceability matters alongside model selection.<\/p>\n

Related projects<\/h2>\n