An Every consulting lead's Codex field notes: how a non-technical operator builds and maintains her own personal work systems
- Natalia Quintero, who leads Every's consulting business, has a non-technical background. Over the past month or two she's moved from Claude Code to OpenAI Codex, using it to build and maintain several personal work systems.
- Her favorite move so far: give Codex a goal — clean up hundreds of customer records in her CRM based on calls and emails — then go to sleep. Six hours later, she woke up and the job was done.
- She built 3 custom apps in Codex: a tool that auto-converts study material into comic strips she can read on her commute; an email triage tool for her own inbox with 7 action buttons that drafts replies in her own voice; and a caregiving coordination app for her 81-year-old father, built in 13 hours.
- Even though she's capable of building her own CRM with AI, she deliberately walked away from the one she'd hacked together and bought Attio instead: software is the skeleton, the AI model is the brain and ligaments — maintaining data quality is literally a CRM company's entire business.
- A year into running Claudia, Every's internal AI employee, the team found it still needs constant human oversight and direction, so they hired an additional human operator.
What she gained by switching from Claude Code to Codex
Natalia Quintero, who leads Every's consulting business, recently went on the AI & I podcast to demo the personal work systems she's built and maintained entirely with OpenAI Codex over the past month or two.
She used to be a heavy Claude Code user, comfortable in the folder structure she'd built herself. Dan Shipper spent weeks nagging her to install Codex; her verdict after switching: the tool changed, and life changed along with it.
For someone who doesn't write code, two changes stood out. First, Codex folds the terminal and browserThe terminal is the black command-line window that makes the machine do things; the browser is the web window. Both are embedded in the same chat interface straight into the chat interface, and paired with a strong, hard-grinding model (5.5), you can feel the compute leaning into hard problems. Second, the burden got lighter: with Claude Code, she used to have to manually plan out the folder structure first, keeping Finder open to watch where things were saved, before AI would work inside that structure. With Codex, the structure grows as it works, so she doesn't have to lay the foundation in her head first.
Running through the whole conversation is a mental-model shift she and Dan keep circling back to: knowledge work is moving from "sculpting" to "gardening."
Sculpting means every outcome is carved by your own hands; gardening means you set up the soil, light, and water so the crop grows itself — you don't tend every single plant by hand. Applied to work: her job is no longer writing every email or updating every record by hand — it's building a system that produces those outcomes automatically, stepping in only at a few key points.
Give it a goal, sleep on it, come back to a finished job
Her favorite move so far happened while setting up Attio (a CRM — customer relationship management software). Her team needed to fill in each customer's record accurately based on what had actually happened in calls and emails.
