Valisthea Psychiatry is fictional. These four figures are from a real US psychiatry practice where I ran operations.
Weeks to about an hour for a pay run, once I wrote it in code. It now runs twice a month.
30-40% less manual admin work, estimated, after I rolled out AI tools for documentation, intake and analytics.
45% to 80-90% collection rate, after I redesigned billing follow-up.
130x patient volume growth in under 2 years, while I ran operations.
01Weekstoaboutanhour
for a pay run, once I wrote it in code. It now runs twice a month.
0230-40%less
manual admin work, estimated, after I rolled out AI tools for documentation, intake and analytics.
0345%to80-90%
collection rate, after I redesigned billing follow-up.
04130x
patient volume growth in under 2 years, while I ran operations.
for a pay run, once I wrote it in code. It now runs twice a month.
manual admin work, estimated, after I rolled out AI tools for documentation, intake and analytics.
collection rate, after I redesigned billing follow-up.
patient volume growth in under 2 years, while I ran operations.
01 / 04
The team
Valisthea OS
Valisthea Psychiatry is fictional and every figure is from my own testing.
The layer that runs all the automations.
Each automation works on its own, but nothing watches them together. So I built a small team of AI employees, each with one job. Shown with sample data.
AI employee
Ivy
Writes the morning brief from every automation's logs: what ran, what failed, what needs a decision.
AI employee
Nora
Answers staff policy questions with the exact quote from the policy.
AI employee
June, chief of staff
Checks their work every Friday and writes a lesson when one gets something wrong.
Half the time the fix is a better process, not another tool. I map it first and tell you which one you need.
02
I ran payroll by hand before I automated it. A pay run that took weeks now takes about an hour. So I build for the bad day, not the demo: no double counting, failures that alert a person, bad data turned away with a reason.
03
You get the documentation and the logins. Nothing only I can fix.
Model + harness + environment
The model is a third of the job.
Most people pick a model and stop there. What you get out of AI depends on three parts working together, and I spend most of my time on the last two.
Model
The LLM doing the thinking. Claude, GPT or Gemini, picked for the job, not out of loyalty.
Harness
Everything wrapped around it: the tools it can call, the steps it follows, what it remembers, and the checks it has to pass. Mine use a second model to review the first, and test cases.
Environment
What it can see and touch: your data, your rules and the apps it connects to, set up so mistakes are hard to make and easy to catch before they reach a customer.
Environment
Harness
Model
Services
Front office
A reply within seconds, day or night
A Harvard Business Review study found a lead answered within an hour is nearly 7 times as likely to turn into a real conversation as one answered an hour later. Mine get a reply within seconds, day or night. Missed calls get a text back, reminders cut no-shows, and patients who drifted away get invited back. For the phone itself, I wrote the setup and the test plan for a clinic's AI phone agent.
GoHighLevel
AI with a person in charge
An estimated 30-40% less manual admin work
That was the drop when I rolled out AI tools for documentation, intake and analytics at a US practice. AI drafts the reply or answers the routine question. A person approves anything that leaves the building.
n8n
Claude
ChatGPT
Gemini
Grok
DeepSeek
Kimi
Qwen
When no-code hits a wall
A pay run that took weeks now takes about an hour
Contractor pay with different rates, payers and splits was too much for a no-code tool, so I wrote it in code. I checked it against past pay runs before it went live, and it now runs twice a month.
Claude Code
Codex
Back office
Collection rate up from 45% to 80-90%
That came from redesigning billing follow-up at a US practice. The builds match payments to visits, check contractor pay before it goes out, and clean intake data before it reaches your team.
n8n
Make
Zapier
01 / 04
Work
Sixteen builds across four tools.
Valisthea Psychiatry is a fictional practice, so every name and number here is synthetic. I built 16 automations for it, then a layer that runs them. Each card has the problem and a result from my own testing.
GoHighLevelF1
Speed to Lead
Sample data, fictional
4 for 4
test inquiries became pipeline cards
The problem
Every new inquiry needs a fast reply and an owner, even from people the practice can't see.
The result
4 test inquiries made 4 pipeline cards, assigned in turn. A form sent twice made one contact.
I'm a licensed pharmacist who moved into operations. I've led operations, finance and logistics teams, and I'm on the leadership team of a US psychiatry practice that grew from a handful of appointments to more than 1,800 a month in under two years.
I build my own tools too. I run a team of AI agents I made in Claude Code that handles my daily brief, sorts my inbox and checks payroll, each with a reviewer and test cases.
I'm based in the Philippines and work Pacific time.
7years
of double-digit growth at a Philippine supply and logistics business I've helped run since 2017
25‑50%
supplier cost cuts I negotiated for that business
Tools
Automation
n8n, Make, Zapier, GoHighLevel
AI
Claude Code, Codex, Claude, ChatGPT, Gemini, Grok, DeepSeek, Kimi, Qwen
Video and visuals
HyperFrames, HeyGen, Higgsfield
Building
Lovable, GitHub, Cloudflare
Knowledge and work
Obsidian, Google Workspace
Training
HighLevel CRM, Tara AI Community, 2026
AI Automation with n8n, Technical Virtual Assistants PH, 2026
Prompt Engineering, Tara AI Community, 2026
Make.com, Technical Virtual Assistants PH, 2026
Zapier, Technical Virtual Assistants PH, 2026
HIPAA Awareness, HIPAATraining.com, 2025
Questions
Who is Rudolph Alejandria?
Rudolph Alejandria is a licensed pharmacist who moved into operations. He has led operations, finance and logistics teams and is on the leadership team of a US psychiatry practice. He runs No Yes Men, is based in the Philippines and works Pacific time.
What is No Yes Men?
No Yes Men is Rudolph Alejandria's AI and automation practice. It builds automations and AI tools for businesses that want to do more without hiring more, and it says plainly what won't work before anything gets built.
What does No Yes Men automate?
The work people do by hand every day: the front desk, follow-ups, billing, payroll checks, reports and inboxes. Most of the work so far has been in US healthcare operations.
What tools does it use?
GoHighLevel, n8n, Make, Zapier and Claude Code, picked to fit the tools a business already has.
How do I start?
Book a call. You get an honest read on what is worth automating and what isn't, before any money is spent.
Tell me what's eating your team's week.
30 minutes on Zoom. You'll leave with a straight answer on what's worth automating, even if it's nothing.
