Google Senior AI Product Manager: 6 agents handle my daily tasks, costing less than $400 per month, running 24/7

Author: Shubham Saboo

Translation: Shen Chao TechFlow

Shen Chao Guide: Six specialized Agents complete research, content creation, code review, and newsletter production while I sleep.

The author fully discloses the file structure, actual costs, pitfalls encountered, and weekly recommendations, making this one of the most practically valuable personal AI Agent implementation records to date.

Full text below:

Six AI Agents handle all my work while I sleep.

This is not a demo. Not a weekend project.

A truly 24/7 operational team ensures I never fall behind. Research is done, drafts are prepared, code is reviewed, newsletters are ready. Every morning when I open Telegram, they’ve already completed a full shift.

Yesterday, I published an article about my Agent team. The top question was: “How exactly did I build this system?”

This is the answer. No theory, no architecture diagrams. I show the actual file structure I use, the real costs I paid, the pitfalls I encountered—all laid out.

After reading this, you’ll understand how to build a self-operating AI Agent team that works while you sleep.

Why a team instead of tools

Managing both Unwind AI and Awesome LLM Apps repositories daily involves six tasks: tracking AI industry hot topics, tweeting, posting on LinkedIn, drafting newsletters, reviewing GitHub contributions, and handling community issues.

Each task takes 30 to 60 minutes. Six tasks. My whole day is gone before I even start real work.

I tried solving this with a single Agent—one massive prompt handling research, writing, and review. The result? Mediocrity. The context filled up, quality declined. One Agent can’t handle six jobs simultaneously.

So I hired six AI Agents.

Meet this team

Each Agent is named after a TV character. Not a gimmick. When I tell Claude, “You have Dwight Schrute’s energy,” it already knows what that means from training data: thorough, focused, treating work as life. This comes from 30 seasons of character traits I’ve freely borrowed.

1. Monica (Chief of Staff): Named after Monica Geller. She’s the main Agent, the one I interact with most on Telegram. She coordinates others, handles strategic decisions, assigns tasks to specialists. Her SOUL.md states: “You are the one ensuring everything is done right.”

2. Dwight (Research): Named after Dwight Schrute. Performs research scans three times daily, checking X, Hacker News, GitHub trending, Google AI blogs, and papers, then writes structured intelligence reports for all other Agents.

3. Kelly (X/Twitter): Named after Kelly Kapoor. Reads Dwight’s research, drafts tweets in my tone, including single tweets, threads, and quote tweets. Her SOUL.md states: “You knew it would trend before it became popular.”

4. Rachel (LinkedIn): Named after Rachel Green. Uses the same intelligence sources as Kelly, but on a different platform and with a different tone—focused on thought leadership rather than hot takes.

5. Ross (Engineering): Named after Ross Geller. Handles code review, bug fixing, and technical implementation. His SOUL.md states: “When solving problems, understand thoroughly first. Don’t just fix symptoms.”

6. Pam (Newsletter): Named after Pam Beesly. Organizes Dwight’s daily intelligence into newsletter summaries.

Each Agent has a distinct role, with clear division of labor.

Now, about building this system

I run everything on a Mac Mini M4. But I must clarify: you don’t need a Mac Mini.

OpenClaw supports macOS, Linux, and Windows (via WSL). Laptops work, gaming PCs work, $5/month VPS works. The Mac Mini’s advantage is always-on, quiet, energy-efficient, but it’s not necessary.

My setup: Basic Mac Mini M4. Always plugged in and connected, no monitor, interacting entirely via Telegram on my phone.

Installing OpenClaw

Just two terminal commands, under five minutes.

If issues arise, consult the OpenClaw documentation.

This launches the gateway, the background process that keeps everything running. It manages your Agents, runs cron jobs, handles Telegram messages. Close the terminal, Agents keep working.

Workspace structure

One OpenClaw instance, multiple Agents—not six separate installs.

My actual directory structure looks like this:

Monica lives in the root directory. She’s the main Agent I interact with directly. Other Agents are her delegated sub-Agents or run independently on their own cron schedules.

You don’t need to start with six Agents. I began with just Monica, and as my workflow became clearer, I gradually added others over several weeks.

What is SOUL.md?

Each Agent is defined by a file: SOUL.md. This is the Agent’s identity, role, and operational instructions—its core.

For example, Dwight’s SOUL.md roughly looks like this:

Pay attention to what this file does. It’s not just saying “You are a research Agent.” It endows the Agent with personality, clear principles, explicit relationships with other Agents, and a decision-making framework.

Monica’s SOUL.md is similar.

All Agents follow the same pattern: identity, role, principles, relationships, style. Each SOUL.md is about 40-60 lines—short enough to fit into context each session, detailed enough to produce consistent, stable behavior.

Multi-Agent coordination

Agents do not call APIs or message queues or use orchestration frameworks.

It’s all files.

Dwight completes research and writes results into intel/DAILY-INTEL.md. Kelly wakes up, reads that file, and drafts tweets accordingly. Rachel reads the same file and drafts LinkedIn posts. Pam reads it and compiles the newsletter.

Coordination is through the filesystem.

Dwight’s SOUL.md precisely instructs him where to write:

Kelly’s AGENTS.md precisely instructs her where to read:

No middleware, no integration layer. Dwight writes a file, Kelly reads a file—handover is a markdown document on disk.

It sounds too simple. It is. That’s why it works. Files don’t crash, no authentication issues, no API rate limits. They just sit there.

Structured data in JSON, human-readable summaries in markdown. Agents read markdown; JSON is the true source of deduplicated, long-term tracking.

Memory system

Agents have no persistent memory of previous sessions; each conversation starts fresh. This is a feature, not a flaw. But it means memory must be explicit.

Two layers.

Daily logs (memory/YYYY-MM-DD.md): raw records of each session, including what happened, drafts created, feedback received. Agents keep writing into these throughout the day.

Long-term memory (MEMORY.md): distilled insights from daily logs—lessons learned, preferences discovered, patterns noticed.

Each Agent, at the start of a session, follows instructions like in AGENTS.md: read SOUL.md, then USER.md, then today’s and yesterday’s memory files, and if it’s a main session, also read MEMORY.md.

These Agents do improve over time—not because the models improve, but because their loaded context becomes richer.

Kelly learns my writing style, no emojis or hashtags needed. That’s stored in her memory, so every draft reflects it without me saying a word. Dwight learns which stories pass the “Alex filter” (our target audience profile) and which to skip—also stored in his memory.

During each heartbeat, Agents periodically review daily logs, extract key content into MEMORY.md. Daily files are raw records; MEMORY.md is distilled wisdom.

Scheduling

Agents need to wake up autonomously. OpenClaw handles this via built-in cron scheduling.

My actual schedule:

Order matters. Dwight runs first, because others depend on his output. Kelly and Rachel run after him, since they need his intel files to draft content.

Heartbeat self-healing

Cron jobs sometimes fail—machine reboots, tasks hang, network drops during API calls. Infrastructure failure modes.

HEARTBEAT.md provides a safety net. Each heartbeat, the main Agent verifies whether cron tasks actually executed:

If a task failed or missed its window, the heartbeat detects it and forces a rerun. Self-healing, no manual intervention.

Heartbeat is suitable for batch checks or scenarios with slight timing drift. Cron is better for precise scheduling and isolated main sessions.

Telegram as interface

No dashboards, no web UI, no admin backend. I interact with Agents via Telegram.

This is deliberate. I don’t want to log into dashboards or web apps. My phone is always nearby, Telegram is always open, and Agents find me where I am.

OpenClaw supports Telegram as a channel. After setup, your Agent appears as a Telegram bot. You send messages, it replies, drafts are sent for approval. Like having a colleague in your chat app.

Monica is my primary contact, handling most conversations and delegating tasks. Other Agents contact me directly when their cron outputs are ready for review.

My typical morning: wake up, open Telegram, Dwight has already sent a research summary, Kelly has three tweet drafts awaiting approval, Rachel has a LinkedIn post ready. I review, give feedback, approve—all in 10 minutes with coffee.

Personality shaping

You don’t start with a perfect personality. You begin with a rough outline in SOUL.md, observe Agent behavior, and correct over time—just like managing real people.

I call this “corrective prompt engineering.”

Kelly’s initial drafts were full of emojis and exclamation points—that’s not my style. I gave feedback: “No emojis, no hashtags, short and punchy sentences.” She updated her memory, and after a week, she was doing it right. Dwight initially captured too much noise—every trending repo, every small update. I told him: “Not everything trending is important. I need signals, not noise.” He updated his principles, and now his reports are focused and actionable.

Any Agent’s first version is mediocre; the tenth is decent; the thirtieth is excellent. You must iterate and refine. Naming them after TV characters gives the model an instant personality baseline—“Dwight Schrute energy” means thorough, focused, no nonsense. But true personality emerges from weeks of stored corrections in memory.

A recommended approach: assign each Agent a simple, mundane job title and a stop condition. Constraints improve performance; the more specific the role, the better the output.

Security

Security is in your hands. My approach is simple: Agents have their own worlds, don’t access my personal accounts.

Mac Mini is their computer. They have separate email accounts, API keys, limited permissions—nothing on that machine connects to my personal accounts.

API keys for services like Gemini, Eleven Labs are dedicated to this OpenClaw instance. I monitor usage, and if anomalies appear, I can cut access within seconds.

I never give Agents access to my personal accounts. If I want them to see an email, I forward it. If I want them to review a document, I share via Telegram. They only see what I want them to see.

This is the same principle as onboarding new employees: don’t give them all keys on day one. Provide their own workspace, credentials, and share information as needed.

Where can problems occur, and how to fix

This is not magic; it’s infrastructure. Infrastructure can fail.

Gateway crashes. Rare but possible. Fix: run “openclaw gateway restart.” The heartbeat system detects outdated cron tasks and forces reruns, so no work is lost.

Cron misses window. Machine sleeps, network drops, API rate limits. Fix: HEARTBEAT.md’s self-healing mode. Monica checks each heartbeat whether tasks actually ran; if any haven’t updated in over 26 hours, she forces a rerun.

Context window overflow. Agent reads too many files at session start, leaving no room for actual work. Fix: keep SOUL.md concise (40-60 lines), focus AGENTS.md, load only today’s and yesterday’s memory files. Agents don’t need to read all history every time.

Output quality decline. Happens when memory files become cluttered or contradictory. Fix: regular memory maintenance. During heartbeat, Agents review daily logs, distill key insights into MEMORY.md, delete or archive old daily files.

Coordination conflicts. Two Agents try to update the same file. Fix: design file flow as “one writer, multiple readers.” Dwight writes DAILY-INTEL.md; others read it. No one writes to it except Dwight.

The biggest lesson for reliability: start simple. One Agent, one task, one schedule. Make it stable for a week, then add a second. Those who deploy six Agents on day one and wonder why it breaks are making the same mistake as deploying a distributed system without monitoring.

Actual costs

Hardware: Mac Mini M4 starts at $499, but any always-on computer works—old laptops, $5/month VPS, whatever you have.

AI model costs: I use multiple models across the team—most tasks with Claude Opus and Sonnet, some with Gemini, and testing local models via Ollama to cut costs further.

Details:

Claude (Max plan): $200/month

Gemini API: $50–$70/month

TinyFish (web Agent): about $50/month

Eleven Labs (voice): about $50/month

Telegram: free

OpenClaw: open source, free

Total: under $400/month for a team that never sleeps.

What has truly changed

Dwight saves me 2–3 hours of research daily. Instead of manually checking X, Hacker News, GitHub trending, and AI blogs every morning, I wake up to a prioritized summary with sources and action items.

Kelly, Pam, and Rachel save 1–2 hours on content drafting. Ross handles engineering tasks I’d otherwise do at night.

Total: about 4–5 hours saved daily.

But the real value isn’t in a single day; it’s in weeks and months of continuous operation. An Agent that researches daily for 30 days accumulates a trackable set of signals, trends, and patterns—something no single session can produce. My posting frequency on X has increased, quality improved, and posting times stabilized. The Awesome LLM Apps repository keeps growing, and the newsletter has a reliable research pipeline.

These Agents can’t do original thinking, strategic shifts, or creative breakthroughs. They handle repetitive, structured work that I used to spend hours on, freeing me to focus on tasks that truly require human intelligence.

How to get started

Don’t try to build six Agents on day one.

Week 1: One Agent, one task. Install OpenClaw, write your first SOUL.md through interaction, pick your most repetitive daily task (research or content drafting for most), set up Telegram, create a cron job, and observe it run for a week, fixing issues as they arise.

Week 2: Add memory mechanisms, keep refining. Your Agent’s initial output will be mediocre—that’s normal. Give feedback, watch the memory files grow, adjust SOUL.md based on what you see. By the end of week two, your Agent should produce genuinely useful results.

Week 3: Add a second Agent. You’ll realize the need—research produces intelligence, but you’re still manually writing tweets. It’s time for a content Agent. Use shared files: first Agent writes, second reads. Coordination via filesystem.

Week 4 and beyond: build sequentially. When you feel the pull, add new Agents—not when you think you “should.” Each should address a real, existing gap in your workflow—not demos or proofs of concept.

Treat it like hiring. You wouldn’t hire six employees on day one; you hire one, get them running efficiently, then add more as workload demands.

Mindset shift

After your Agents run stably for a month, something changes. You no longer see AI as a tool you open when needed, but as a team that’s always working.

I start mornings greeting Monica with “Good morning,” and before I turn off my phone, I say “Good night” to the team. It sounds absurd, but after a month of daily interactions, feedback, and seeing their progress, the line between Agent and human begins to blur.

Models are the baseline—everyone can use Claude, GPT, Gemini. The “system”—SOUL.md, memory, scheduling, coordination, weekly feedback stored in files—is your unique setup.

That system is yours. No one else has your Agents, your memory files, your personality tuning.

And it compounds daily.

Dwight’s scans enrich his memory; Kelly’s feedback sharpens her drafts; Ross’s bug fixes deepen his understanding of your codebase.

That’s the real moat—not the models, but the learning system.

Start today. One Agent, one task, one schedule.