Hector Herrera · AI systems architect
Systems stories, not portfolio thumbnails.
Most of my work now is AI implementation for working teams: internal AI help desks, knowledge bases, intake questionnaires, follow-up automation and the dashboards that show it all working. Each story covers what the problem was, which systems were involved, what I owned, what we decided and what changed.
What I’ve built and run
-
Main business · Founder
Hex AI Systems
A proprietary AI operating system that sits on top of today’s leading AI platforms and turns a business’s workflows into agent-driven operations: a centralized brain that gets smarter every session.
I founded the company and built the framework it runs on.
Visit hexaisystems.com -
Studios of Hex AI Systems
-
Studio · Founder
HEX Web Design
Fast, easy-to-manage websites for Houston businesses, wired to the forms, tracking and follow-up that turn a visit into a customer.
Every form is tested to prove it reaches the client before launch.
Visit hexwebdesign.com -
Studio · Founder
HEX Creative Works
Social media for Houston businesses. Strategy, content, posting, community and paid social.
Every post is made for the buyer, and a person checks it before it goes out.
Visit hexcreativeworks.com
-
-
Side project · Creator & publisher
NexChron
Daily AI intelligence: how AI is reshaping business, government and everyday life.
I built the site, the publishing pipeline and the daily briefing, and I write under my own byline.
Visit nexchron.com
Fractional AI officer
Some businesses need someone to own AI strategy and delivery without hiring a full-time executive for it. For those, I take on a small number of fractional AI officer roles.
That usually means setting the AI roadmap with leadership, deciding where AI goes first, putting the knowledge base and review habits in place, and building or overseeing the systems themselves. It's the same work described on this page, with a seat at the leadership table and the accountability that comes with it.
Engagements run through Hex AI Systems, and the full outline of how they work is on its site.
The kinds of problems I work on
AI operations for real teams
Assistants that take requests from team chat and email, sort them, draft replies and bring the decisions to a person. Anything that can change a live system runs behind safeguards and an undo.
Knowledge systems AI can work from
The rules, roles, lessons and verified facts an AI needs to act like it knows the business, organized so each task reaches the right expertise and every correction sticks.
Intake and delivery platforms
Multi-step intake with screening, e-signature, questionnaires, document upload and follow-up. The business owns the record, and every submission is proven to reach a person.
Monitoring that catches silence
Checks that notice when something expected didn’t happen. No leads when there should be some, a reply that never went out, a report that quietly stopped updating.
Search and AI visibility
Making sites authoritative in Google and inside AI answers through stronger content, entity signals, structured data and citations, then tracking whether the engines actually cite you.
Analytics, attribution and first-party data
Following a click from the ad to the CRM and back, and owning the data instead of renting visibility from vendor dashboards.
Multi-site portfolios
Running many sites across different stacks and CRMs with repeatable deployment and one ledger of what exists, where it lives and who owns it.
Documentation that stays true
Owner’s manuals written from the running system, versioned with every change, with checks that catch the day the book and the system disagree.
Infrastructure decisions
Hosting, caching, mail and security on servers I run myself, and honest tradeoffs between heavy platforms and leaner custom builds.
Work you can inspect today
Client work comes with confidentiality. The stack I run myself doesn’t, so it’s the clearest look at how I build: the server, mail, automation and this site.