Email open rate after building a newsletter system from scratch.
(Industry average:
20 to 25%)
AI & Automation Specialist
Building systems that keep running when you’re not in the room
I'm May. Before I build anything, I map who does what, why they do it that way and what would break if I changed it.
Results
Some outcomes from systems I’ve worked on
This usually comes from changing how things are set up.
Minors completed a full GDPR-compliant enrollment flow, parental consent, double opt-in and platform access, without a single manual review.
From a $130K investment, after rebuilding enrollment and onboarding end-to-end
(823% ROI - System designed and built by myself)
Selected work
Three systems, real constraints and nothing that only worked in testing
Each one started by understanding the context first.
700+ minors processed without manual review, GDPR-compliant access control built from scratch in a month
A policy-driven system where access couldn't depend on trust or manual checks, so each step was designed as a gate. If a condition wasn’t met, the system stopped progression by design.
Manual operations cut from 80% to 20%, onboarding rebuilt so every step triggers the next
Onboarding worked, but only if everyone remembered what to do. At scale, that stopped being reliable. In this case, the flow was redesigned so each step triggered the next automatically.
$1.2M from a $130K investment, once tracking, CRM, email and payments finally worked as one system.
Campaigns were running, but the data couldn't be trusted. Tracking, emails, payments and CRM were each doing their own thing. Everything was rebuilt as one system so performance could be traced, repeated, and scaled without starting from scratch every time.
Before I build anything
How I work
01
Map the people
Who actually does the work and not just who manages it. The real process lives with the people doing it.
02
Understand the why
Every process has a reason, even the ones that look broken. That's why before changing anything I need to know why it works the way it does.
03
Find the edges
Edge cases aren't exceptions. I map what breaks, who catches it, and what happens when nobody does.
04
Then build
Now the system has a chance of being actually used, designed around how the team already works.
Want to know more?
Reach out on LinkedInAbout
Build around context. Automate to scale.
I come from Fine Arts and now I design AI operational systems.
These two things are more related than they look, because both require figuring out how things should flow, where they break, and why the instructions nobody reads are always the wrong ones.
The path here wasn’t obvious. It came from connecting things that didn’t seem related at first.
Fine Arts led to frontend development at Tuenti (Telefónica), that led to paid media and campaign strategy, which led to automation and which led me finally here. At the end, each step added something the next one needed.
That’s what allows me to understand how the whole company actually operates.
I also hold a Google Cybersecurity Certificate, not because I'm pivoting, but because systems that handle sensitive data need to be designed with failure in mind from day one, not after the first incident.
The tools I actually use the most:
- Make, Zapier and n8n for orchestration.
- Claude API and OpenAI for AI-powered workflows.
- Pipedrive, ActiveCampaign and Airtable on the CRM and ops side.
- APIs, Webhooks and JavaScript when the integration doesn't exist yet.
The $1.2M campaign ran on Make, ActiveCampaign and Pipedrive. The GDPR consent flow ran on Zapier and PandaDoc. That's why tools shift, but the approach doesn't.