Organize your business operations by creating a 'fellowship' of specialized AI agents. This workflow shows how to silo responsibilities into projects and assign distinct roles to AI assistants to automate administrative tasks and context retrieval.
with Claire Vo
I just called him Gandalf. Straight up. And then I asked Gandalf to create everyone else. I needed other bots for engineering, marketing, project, uh, product management, account sales, and customer success... Elrond, the elf, runs engineering. I have Gimli, the dwarf... runs sales. And Sam always plays support.
An operational system where distinct AI agents manage specific business functions, capable of retrieving information and executing tasks with high-level commands.
OpenAI conversational AI
OpenAI's cloud-based AI software engineering agent that can execute code, run tests, and handle complex multi-file tasks autonomously.
All-in-one workspace
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
5 steps
Before automating, you must know what you do. Create an exhaustive list of all your business roles and recurring tasks.
William listed his functions in Notion: customer success, inbound sales, product, accounting, HR, and hiring. This list is the blueprint for your agent system.
In a tool like ChatGPT or Codex, create a separate project for each major business area (e.g., 'Operations,' 'Sales,' 'Growth'). This helps organize context by task; check your assistant’s actual memory and project settings before assuming isolation.
Think of each project as a dedicated workspace for a specific department. All conversations about sales happen in the 'Sales' project.
Create distinct personas for agents that correspond to business functions. Start by creating a 'chief of staff' agent and instructing it to define the others.
You are Gandalf, my chief of staff. Your first task is to define a fellowship of AI agents to help me run my business. Create personas for agents responsible for Engineering, Marketing, Sales, Accounts, and Customer Success, and describe their primary duties.
The names (Gandalf, Elrond, Frodo) make the system memorable, but the key is the operational separation of duties.
Connect your agents to your company's systems of record, such as Google Suite, Notion, Linear, and QuickBooks. This allows them to retrieve information autonomously.
The source recommends extreme caution here. Unlike William's 'allow always' approach, a safer path is to start with read-only access and enable specific write permissions with human oversight.
Test the system by giving a high-level command that requires the agent to access multiple sources and perform an action.
I just finished a call with [Candidate Name]. Find their information in my email and calendar, create a profile for them in our 'Three Pipes' recruiting tool, and give me a summary.
Success is when the agent can execute the task without you having to manually copy and paste information between applications. However, always verify the output for accuracy.
Turn an idea into a PRD, user stories, and a plan.
How to recover when the loop fails and where human judgment helps.
Let routine work keep moving.
Build a voice-first language practice app that forces immersion by refusing to switch to the user's native language. The AI tutor provides corrections in real-time, within the flow of the conversation, to simulate an authentic learning environment.
Use Codex to help build a native mobile app, then guide the design and review the resulting code. William brought web-development and product experience to the project.
Create a repository-based system to provide an AI assistant with durable context about your team, projects, and communication channels. This enables it to perform routine tasks like summarizing daily updates and identifying stakeholders, acting as a personalized chief of staff.
Join 250,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.
