Seth Nichols

Agriculture. Business.
Technology.

I help producers make better decisions — with clear information, honest financials, and systems that match how the operation actually runs.

Read my thinkingMy story

CFO · Arizona Dairy Co   |   Building HerdOptix

Philosophy

What I keep coming back to.

Systems create better decisions

Good people without structure can still fail. Clear processes, good information, and accountability give people a real chance to succeed.

Data informs judgment. It doesn't replace it.

I want the data first. But experience and instinct still count until the evidence says otherwise — there's often more to the story than the spreadsheet.

Fix causes, not symptoms

Patching the same problem every week isn't maintenance — it's avoidance. We don't need Band-Aids. We need stitches.

Simple usually wins

Technology should remove complexity, not add another layer of it. If a tool makes the operation harder to understand, it's the wrong tool.

Control the controllables

Understand what a person or a business can actually influence — and hold them accountable for that, nothing else.

Leadership matters

Underdeveloped leaders create underperforming organizations. Sometimes the employee isn't the problem — they're in the wrong position, or inside a bad system.

Pride shows up in the details

You can learn a lot about a business before you ever see its financial statements. Cleanliness. Maintenance. How the equipment is organized. How ownership treats the people doing the work. A great operation runs like a Swiss watch — everything has a place, people understand the standard, and there's pride in the operation.

How I Work

Data → Financials → Operations → Decisions → People.

Every complicated business problem walks the same path: start with what is actually happening, and end with people equipped to execute. The order matters.

01

Data

Start with what is actually happening — measured, not remembered. Get the numbers before the opinions.

02

Financials

Translate it into dollars. Cash flow, margin, risk — the financials are where operational reality eventually shows up.

03

Operations

Then go stand in the operation and understand why the numbers look the way they do. The spreadsheet rarely tells the whole story.

04

Decisions

Make the call. A good decision made on time beats a perfect one made too late.

05

People

Give people the structure, direction, and accountability to execute it. Systems don't run themselves.

Story

It starts in Gilbert.

Not with a company or a pitch — with helicopters, cattle, and a database nobody asked me to build. I'm still learning. That's part of it.

Gilbert, Arizona

I grew up a quiet kid who learned fast and didn't particularly enjoy school. What I loved was aviation. I started flight training at 16 and earned my helicopter certificate at 17.

Cattle

Then cattle. I went to work for a cattle company where my dad worked. I was interested in databases, so I built a cattle inventory system. Gradually, that made me the inventory guy.

Stepping up

When the owner died, I had to step up. I moved deeper into the finance side and eventually carried significant financial responsibility in cattle businesses.

Hard lessons

Then came a period I'll describe simply: business failures and hard lessons. It changed how I think about risk, controls, trust, information, and systems.

2021 — Dairy

I entered the dairy business with essentially one assignment: help us become profitable. That turned into relocating a dairy, taking the herd from roughly 7,000 cows to 3,700, rebuilding a run-down facility, cash-flow forecasting, CAPEX decisions, financing, reporting, and operations.

HerdOptix

HerdOptix started almost laughably small: a report everyone hated producing. Thirty minutes became ten. Ten became two. Then I asked why anyone should have to produce it at all. Five years later, that question has become a software company.

Projects

Things I've built.

Problems I found interesting enough to build a system for.

HerdOptix Platform

Dairy & Ag

Problem

Compliance records, workforce management, genetics, and herd data lived in separate tools and spreadsheets — nothing talked to each other.

Solution

A unified dairy operations platform — compliance, workforce, genetics, and an AI agent that answers questions across all of it — built on one Supabase data layer.

Outcome

One system of record for the whole operation. Reporting that took hours is now a question typed to the agent.

Next.jsSupabaseTypeScriptClaude API

Vector Solutions Agents

AI Agents

Problem

Business owners want an AI assistant working their actual email and Slack — not a chatbot in another browser tab, and not shared infrastructure they can't trust.

Solution

Self-serve platform that provisions an isolated AI agent per customer — connected to email, Slack, and iMessage — each running in its own sandboxed environment.

Outcome

Every customer gets a dedicated agent in their own channels, with hard isolation between tenants — no shared context, no crossover.

TypeScriptNext.jsFly.ioClaude API

ETL Pipeline Suite

Data Infrastructure

Problem

Herd software, milking parlor databases, PLC datalogs, scale tickets, and market prices all lived in separate systems with no common format.

Solution

Scheduled ingestion services — API syncs, database pulls, even reading the barn PLC directly — that normalize everything into one clean Postgres layer.

Outcome

Hundreds of data points ingested daily without manual intervention. Dashboards and the AI agent run on current, unified data.

Node.jsPythonPostgreSQLRailway

Farm Scale Kiosk

Automation

Problem

Truck weights were captured in aging desktop scale software — one PC, manual tickets, and data nobody else could see.

Solution

Self-serve kiosk at the scale with a touch UI, a sync service that pulls data out of the legacy scale database, and a web portal for weights and tickets.

Outcome

Drivers weigh in without an operator, and scale data lands in the same platform as the rest of the operation.

Next.jsTypeScriptDockerSQL Server

Maintenance QR System

Automation

Problem

Field teams had no way to log equipment maintenance without going back to a desk. Paper logs got lost or were never filed.

Solution

QR codes on every piece of equipment link to a mobile-first form — scan, log, done. No login, no app install, no friction.

Outcome

Eliminated paper logs entirely. Maintenance history is searchable and tied to specific equipment.

Next.jsSupabaseTypeScript

Operational Alert System

Automation

Problem

Critical events — scale errors, feed variances, reproduction alerts — were discovered hours late because they required checking dashboards.

Solution

Event-driven alert pipeline that detects anomalies and routes notifications to phone via ntfy. Configurable thresholds per event type.

Outcome

Immediate visibility into operational issues. Problems caught in minutes, not hours.

GoPythonntfyPostgreSQL

HerdOptix

Related brands.
Not the same brand.

My thesis

Help producers make better decisions.

HerdOptix's thesis

Give producers the systems and intelligence to do it.

Non-negotiable

HerdOptix sits on the producer's side of the table. The producer owns their raw data — the company earns its keep on what it builds on top: normalization, intelligence, benchmarking, and models. That principle matters more than squeezing every possible dollar out of the platform.

herdoptix.com →

Ideas I'm Exploring

What I'm thinking about building next.

Cattle Auction Analytics

Price prediction and historical trend analysis for cattle auctions. Know what to bid before you walk in the ring.

Embryo Marketplace

Centralized platform for buying and selling embryos with verified genetics, health records, and custody chain.

Agent-Run Back Office

AI agents handling the bookkeeping, inbox triage, and follow-ups for a small operation — plain-text accounting a human can audit, with agents doing the data entry.

Autonomous Research Agents

Point an AI agent at a real experiment loop and let it run overnight — change the code, test, keep what works, repeat. Wake up to a log of experiments and a better result.

Blog

Writing.

All posts →

March 26, 2026

The Pilot Trap: When Testing Becomes a Way to Avoid Deciding

Perpetual piloting isn't caution — it's a way to spend money on AI without getting any of the benefits, while competitors compound real operational advantages.

March 26, 2026

You Don't Have an AI Problem. You Have a Process Problem.

AI failure is almost never the AI's fault — it's what happens when you drop a powerful tool on top of a process nobody ever wrote down.

March 26, 2026

A Practical AI Stack for an Operating Business

You don't need 15 AI tools — you need five, selected for specific use cases, integrated with your existing systems, and actually used.

Life Around It

Not everything is a business lesson.

Aviation, cattle, dogs, travel, the occasional ball game. I've been flying since I was 16 — helicopters first. Sometimes an airplane picture just means it was a beautiful morning to fly.

AviationCattleDairiesDogsTravelBaseball

Contact

If something here resonates,
reach out.

No funnels, no forms. Email works.

seth@sethnichols.comX / @SethHNicholsLinkedIn