I Built an AI Agent Team That Runs My Business
TL;DR
AI agents run on persistent cloud computers: Unlike ChatGPT, these agents operate 24/7 with their own file systems, can download videos, create documents, and maintain state across conversations.
You can message agents like regular contacts: Agents get phone numbers for iMessage/SMS and can be added to Slack channels, making them accessible wherever you already communicate.
The GRASP framework structures agent creation: Goals, Resources, Automations, Skills, and Personality provide a checklist for building agents that operate autonomously.
Agents can create their own skills on demand: Ask an agent to reference your content transcripts and save it as a skill, and it will write scripts in your voice automatically.
Automations run on schedules you define through conversation: The hiring agent was set up with four daily cron jobs (sourcing at 8:30am, outreach, review digest at 4:30pm, nightly reset) just by asking in Slack.
Agents build working tools for you: The hiring agent created a mobile-optimized Trello-style dashboard with candidate cards, fit scores, and status tracking entirely through natural language requests.
The Breakdown
Riley Brown spent over 100 hours building a team of AI agents that operate 24/7 across iMessage and Slack, demonstrating how to deploy a marketing agent from a template in under 20 minutes and build a hiring agent from scratch that sources candidates, manages outreach, and maintains a live hiring dashboard. The agents run on persistent cloud computers, connect to tools like Notion and Linear, and can be controlled entirely through natural language in your existing messaging apps.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
The Cheapest Model That Passes
OpenRouter lists 400 models behind one API. The fix for choosing isn't a better leaderboard, it's a four-step protocol that ends in a real eval.

Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.

Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.