Hey Prompt Warrior,
I added up the manual work my agency was drowning in and built an automation for each one. Together they save me more than 10 hours every week now.
Every single one follows the same pattern: a skill that holds the logic, and a routine that just triggers it. That's the whole method. Once you see it, you can build any of these yourself.
If you prefer watching over reading, here's the YouTube video.
Here's what we'll cover:
Personalized call reminders that lift your show rate without faking a human
Automated proposals built straight from the call transcript
Sales-call reviews scored against the Hormozi CLOSER framework
Client onboarding that spins up folders, pages, and a roadmap in one command
Monthly metrics pulled from Stripe and your CRM into one sheet
The skill + routine pattern that ties them all together
Read time: 6 minutes
📣 Quick plug
If you're an agency owner, coach or consultant doing $10K-$100K/mo, I'd love to help you deploy AI agents across your business.
What you get:
Hours back every week that used to disappear into admin work
Systems that keep running when people leave, built on agents instead of headcount
A real shot at being first in your industry before competitors catch up
Spots are limited. If that sounds like you, book a call here.
Now, back to the tutorial.
🧩 The pattern behind all five
Before the automations, the method, because it's the same every time.
I write a skill first. It's a text file that holds everything Claude needs: what to pull, how to analyze it, where to save the output. This is the part that used to be a tangled n8n workflow.
Then I create a routine, which is just a trigger. Two words: "run [skill name]". The routine never changes. It fires on a schedule and always picks up the latest version of the skill.
The payoff: when I want to improve an automation, I edit the skill. I can also run any skill on demand without waiting for the schedule. Logic and trigger stay separate.
Connecting tools is the only setup step. The fastest way is the MCP connector library inside Claude Code (Google Calendar, Attio, Fathom, Notion all live there). For a few tools I use the CLI or direct API instead, since they're leaner on tokens.
📞 Automation 1: Personalized Call Reminders
I got this from a Hormozi sales training. One of the biggest levers on revenue is your show rate, and a personalized video sent the day before a call moves it a lot.
Best for: any team running discovery or sales calls with a no-show problem.
How it works:
The skill reads today's sales calls from Google Calendar.
It pulls each prospect's details from Attio, then does a quick web check on their business.
It drafts a short, personal video script for each one, plus their phone number so I can drop it straight into WhatsApp.
I record and send the video myself. I keep that part human on purpose. If a prospect can tell it's an AI avatar, you lose the trust the message was supposed to build.
Time saved: about 40 minutes a week at 10 calls. Small on its own. It stacks.
📄 Automation 2: Automated Proposals
Custom proposals eat 30 minutes each. This one builds them from the call.
Best for: turning discovery calls into sent proposals without the writing time.
How it works:
The skill reads the full call transcript from Fathom.
It builds a 5-section proposal: where you are today, your goals, the ROI math, the recommended offer, and next steps.
It stops for my approval (I add a feedback loop on anything client-facing), then writes the proposal into Notion.
It updates the Attio deal from "discovery scheduled" to "proposal sent." An optional step sends it over email or WhatsApp.
I run it with /proposal-builder and the prospect's name. I tuned it to be concise with bullet points after the first version came out way too verbose.
Time saved: roughly 4.5 hours a week at 10 calls.
🎯 Automation 3: Sales-Call Reviews
Most people never go back and study their calls. So they stop getting better at selling.
Best for: coaching yourself or a sales team without daily live review sessions.
How it works:
Every morning the skill pulls yesterday's calls from Fathom.
It scores each one against the Hormozi CLOSER framework, a letter-by-letter scorecard.
It writes a coaching doc into Notion: executive summary, what went well (with real quotes from the call), missed opportunities, key moments, talk ratio, and action items.
I read the output in about 3 minutes over coffee. I've set this up for myself and for clients.
Time saved: around 4.5 hours a week, before you even count the close-rate gains as the calls get better.
📁 Automation 4: Onboarding + Roadmap
A strong first 30 days drives retention. So onboarding should feel sharp, not slow.
Best for: agencies and coaches who run the same setup steps for every new client.
How it works:
I trigger the skill with a client name.
It creates a Drive folder for the client with Sales, Marketing, Operations, and Skills subfolders.
It creates a Notion page under "Clients," then calls my roadmap skill to generate their AI System Roadmap.
It adds a to-do list built from that roadmap and drops the Drive link into the Notion page.
One command replaces a folder-and-doc setup that used to take half an hour.
Time saved: about 10 hours a month at 20 new clients.
📊 Automation 5: Monthly Agency Metrics
Tracking metrics by hand means pulling numbers from Stripe, the CRM, and a dozen other places. So most people skip it.
Best for: anyone who wants a monthly scorecard without the data-gathering slog.
How it works:
Once a month the routine fires the skill.
It pulls deals by stage from Attio and revenue from Stripe, and logs social output (newsletter, Twitter via Typefully, YouTube, TikTok and more).
It confirms the figures with me, then writes them into a Google Sheet: revenue dashboard, entries, sales funnel, and social.
I added a hard rule: never fabricate numbers. The approval step stays until I trust a routine fully.
Time saved: at least 2 hours a month, calculated conservatively.
🧭 The Decision Tree
What you're automating | Tools it connects |
|---|---|
Call reminders | Google Calendar + Attio |
Proposals | Fathom + Notion + CRM |
Call reviews | Fathom + Notion |
Onboarding | Notion + Google Drive |
Monthly metrics | Stripe + Attio + social APIs |
🎯 TLDR
✅ Build a skill for the logic, a routine for the trigger, and keep them separate
✅ Personalized call reminders lift show rate; keep the recording human
✅ Proposals build straight from the Fathom transcript and update your CRM stage
✅ Sales-call reviews score every call on the Hormozi CLOSER framework, every morning
✅ Onboarding spins up folders, a Notion page, and a roadmap from one command
✅ Monthly metrics land in one sheet from Stripe, your CRM, and your socials
Thanks for reading!
If you enjoyed this, consider forwarding this newsletter to a friend or colleague.
What would you like me to write about in the next post? Reply to this email and let me know.
P.S. Wiring these into a real business, connecting the tools, writing the skills, and setting the routines so they run without you, is exactly what I do inside AgentOS. If you'd rather have the system built for you than piece it together yourself, take a look here.
