AI Systems for Small Business

Your bridge from manual work to AI that runs reliably.

PuenteWorks builds multi-agent AI systems, automates workflows, and helps small businesses adopt AI without the chaos. Bilingual English / Español.

What we've shipped
19
tools in one MCP server mcp-video: production AI tooling with 276 automated tests
8k+
people, one ops system Training program operations at enterprise scale
3–5h
saved per week with Copilot Early AI adoption at Capital Group — built the playbook
Simon Gonzalez De Cruz Founder · Long Beach, CA · English / Español
Experience from
Capital Group
SoCal Edison
PIMCO
American Red Cross

AI systems that work in the real world.

Not demos. Not pilots that die in a slide deck. Systems with evaluation, failure handling, and operations baked in from day one.

Multi-Agent Systems

End-to-end agent architectures that decompose complex workflows into reliable, testable sub-tasks. Built with production trust boundaries and failure recovery — not just a chatbot.

Workflow Automation

Automate the high-volume, rule-governed work that eats your team's week: scheduling, data entry, notifications, report generation, ticket triage. Reclaim hours, not minutes.

AI Adoption Strategy

Bilingual (English / Español) consulting on where AI actually helps your business, how to evaluate it honestly, and how to introduce it to a team without losing trust.

Specify. Build. Evaluate. Operate.

Every AI project we take on follows the same discipline. We don't ship something that just looks like it works.

1

Specify the problem precisely

Vague requirements produce vague AI. We decompose your workflow into concrete sub-tasks with explicit inputs, outputs, and edge cases before writing a line of code.

2

Build with trust boundaries

Every action that's hard to reverse gets a human checkpoint. We design for what happens when the AI is wrong — because it will be, sometimes.

3

Evaluate, don't just ship

We build test suites before we deploy. Our mcp-video server shipped with 276 automated tests. Your system gets the same treatment.

4

Operate and improve

We design for monitoring: quality drift detection, failure alerts, and metrics that tell you when the system is working and when it isn't.

The gap isn't the AI. It's the people who know both the domain and the system well enough to make it work reliably.
Simon Gonzalez De Cruz Founder, PuenteWorks
11 years enterprise ops + AI systems builder

Pick your starting point.

Three ways to work together, depending on where you are and what you need right now.

Fast outcome

6-Week AI Pilot

We identify one workflow that costs your team real time, design an AI agent system to automate it, build it, and deliver a working system with documented ROI.

  • Workflow analysis and agent specification
  • Production system with evaluation framework
  • Executive memo with quantified time/cost savings
  • Handoff documentation and your team trained
Start a pilot →
Build for scale

Custom Agent System

A fully designed and built multi-agent system for your specific domain. Multi-step workflows, tool integrations, trust boundaries, monitoring, and failure recovery.

  • Full architecture and specification design
  • Multi-agent build with automated test suite
  • Integration with your existing tools and data
  • Ongoing monitoring and iteration support
Discuss your project →
Advisory

AI Strategy & Enablement

Not sure where to start? We assess your workflows, identify the highest-ROI AI opportunities, and build a roadmap your team can actually follow. Bilingual delivery.

  • Workflow audit and AI opportunity mapping
  • Tool recommendations with honest tradeoffs
  • Team enablement in English or Español
  • Governance checklist and policy templates
Book a strategy call →

Built and shipped. Not slides.

What we point to when someone asks what we've actually done.

276
Automated tests on one MCP server
mcp-video ships with a full test suite. 19 tools for video editing AI agents. Public on GitHub. This is what "production ready" means to us.
8,000+
Associates in one ops system
Managed global training operations at Capital Group — scheduling, enrollment, LMS administration, quality metrics — for 8,000+ people simultaneously.
3–5h
Saved per week with Copilot
Was an early Microsoft 365 Copilot adopter at Capital Group. Identified use cases, built prompt templates, and introduced AI workflows to the whole team.

Simon Gonzalez De Cruz

I build AI agent systems and consult on AI adoption. My background is unusual: 11 years of enterprise operations at Capital Group, Southern California Edison, and PIMCO — then I started building the AI systems myself.

I've shipped a production MCP server (mcp-video, public on GitHub), a multi-agent executive function assistant, a self-recursive creative coding agent, and a Spanish dialect translation MCP server published to npm. Each one taught me something the docs don't cover.

I'm based in Long Beach, California. I work in English and Spanish — fluently in both. If your team or your clients are bilingual, that matters.

PuenteWorks means "bridge works" — building bridges between where you are and where AI can take you, without the hype and without the chaos.

Let's figure out if AI is the right tool for your problem.

First conversation is free. No pitch deck. We'll look at what you're doing, what's costing you time or money, and tell you honestly whether AI helps.

Send a message

Tell me what you're working on. I'll respond within one business day.