A Working System, Then a Plan

V.E.T.S. is not a proposal for a platform. It is a platform, in production, with a proposal attached.

This page puts the evidence first. What the system already contains, what the AI layer already does, and only then what it would cost to finish and how it earns. Everything in the first half can be checked against the database. The figures in the second half are ours to state and yours to question.

What is already built

The V.E.T.S. database holds:

1,113
tables
2,283
views
2,834
stored procedures
249
functions

Scale on its own proves very little. The figure that does the real work is this one: the platform manages 47,062 animal records across its tenants. That is a live count rather than a projection, and it is the difference between a system that runs and a system that demonstrates.

See what 2,834 procedures are actually doing →

The AI layer is deployed, not planned

Fourteen AI assistants, called minions, are configured across 78 page contexts in the application. Behind them sits a retrieval index of 2,333 documents, deployed on Vertex AI Vector Search as vets_deployed_v3.

The architectural decision worth an executive’s attention is duller than either of those numbers. No model name is written into the application. Every model string lives in stbl_System_CodesGlobal and is read at runtime through astp_AI_GetModelConfig: gemini-2.5-flash for interactive chat, gemini-2.5-flash-lite for nightly batch work, gemini-3.1-pro-preview for premium evaluation.

That is why the premium tier moved from gemini-2.5-pro to gemini-3.1-pro-preview as a data change rather than a deployment. In a market where the best available model changes every few months, the cost of switching is a business input, and ours is close to zero.

One boundary worth stating plainly, because cautious buyers ask it first: the vector index holds documentation about the system — help text, knowledge-base articles, and generated descriptions of database objects. Your animal records are never embedded and are not in the index.

How the retrieval and minion layer fits together →

Why there is a market here at all

Physical assets depreciate. A barn, a stock trailer, an ultrasound unit are each worth less every year they are owned. Knowledge assets move the other way. A treatment protocol that forty practitioners have corrected is worth more than it was the day it was written.

Almost no animal-care organization owns its knowledge in a form that can appreciate. It sits in the heads of experienced staff, in paper files, and in the margins of software built for billing. When someone retires, most of it leaves with them.

What fragmented animal knowledge costs →

The commercial consequence is what makes this a market rather than a cause. Software can be copied in a quarter. A corpus that thousands of practitioners have spent years correcting cannot be, and every correction makes the next one more valuable. The defensible asset is the curated knowledge and the community maintaining it.

How expert corrections compound →

$12 Million Five-Year Open Development Plan

Everything above this line can be verified. Nothing below it can. This is a budget, and a budget is a statement of intent. We publish it in full because an open development plan that hides its own numbers is not open.

Block 5-Year
Omega Consulting & Development $6,000,000
Humble Hairpin R&D Facility $4,000,000
Operations & Administration $2,000,000
Total $12,000,000

Omega Consulting & Development — $6,000,000

Development and architecture work delivered against Appframe, the enterprise framework the platform is built on.

Category Monthly 5-Year Purpose
Platform Development $60,000 $3,600,000 Core feature development, AI integration, modernization
Architecture Consulting $25,000 $1,500,000 Enterprise patterns, scalability, security hardening
Support & Maintenance $15,000 $900,000 Bug fixes, updates, production stability

Humble Hairpin R&D Facility — $4,000,000

A working ranch in Texas dedicated to IOT research with horses, cattle and exotics. Animals and husbandry expertise come from the existing ranch operation, which is why the labor line funds instrumentation staff rather than a herd.

Labor — $2,400,000

Role Monthly 5-Year Purpose
Research Director $12,000 $720,000 IOT strategy, vendor relations, protocol development
Field Technicians (2) $16,000 $960,000 Sensor installation, data collection, equipment maintenance
IOT Systems Engineer $7,667 $460,000 On-site technical integration, sensor deployment, data pipelines
Ranch Hands (stipend) $4,333 $260,000 Husbandry and animal handling

Construction — $1,200,000

Category 5-Year Purpose
IOT Lab Buildout $400,000 Sensor workshop, calibration station, network infrastructure
Animal Housing $500,000 Instrumented stalls, paddock sensors, covered arenas
Data Center $300,000 Edge computing, local storage, weather station

Maintenance — $400,000

Category 5-Year Purpose
Equipment $200,000 Sensor replacement, calibration, upgrades
Facilities $150,000 Fencing, structures, utilities
Veterinary Care $50,000 Research animal health, emergency care

Operations & Administration — $2,000,000

Labor — $1,200,000

Role Monthly 5-Year Purpose
Operations Manager $10,000 $600,000 Business operations, vendor management, compliance
Administrative Support $6,000 $360,000 Bookkeeping, scheduling, communications
Part-time Specialists $4,000 $240,000 Legal, accounting, marketing as needed

Construction — $300,000

Category 5-Year Purpose
Office Buildout $200,000 Workspace, meeting facilities, demo area
IT Infrastructure $100,000 Network, servers, security systems

Maintenance — $500,000

Category 5-Year Purpose
Technology $300,000 Cloud services, software licenses, hardware refresh
General Operations $200,000 Insurance, utilities, supplies, travel

Revenue model: four income streams

1. IOT ecosystem

Sensors woven into the environment an animal already lives in, rather than strapped to the animal: water-trough consumption, gate counts, thermal imaging, arena footing, feed-bin levels. Each sensor is both a recurring subscription and a data point that makes the platform better at its main job.

Revenue: hardware sales, installation services, monthly monitoring subscriptions, premium analytics packages.

2. Contextual intelligence advertising

Not banner advertising. A veterinarian opening a protocol and finding they are two doses short is a moment when a supplier’s next-day price is useful information rather than an interruption. The same is true of a breeding-window calculation, or equipment overdue for replacement. The test is whether the reader would have wanted the information without the sponsorship.

Revenue: cost-per-action on purchases and quote requests, premium placement, vendor subscription tiers.

3. The compute economy

Every AI platform has a compute bill that grows with usage. V.E.T.S. treats two kinds of compute as real. Silicon compute is tokens, embeddings and vector queries, measurable in dollars. Biological compute is expert correction, protocol validation and domain judgment, and no quantity of API calls will generate twenty years of clinical experience.

That inverts the usual pricing logic. In most software the heaviest users cost the most and should pay the most. Here, the expert who corrects a shoeing protocol improves a knowledge chunk that answers questions for everyone who asks later. Their expertise subsidizes the platform, so the pricing should recognize it rather than penalize it.

Four tiers follow. Explorers use platform-funded compute at a limited rate. Contributors supply their own model API key, so their usage runs on their own compute. Patrons consent to a share of their compute improving community knowledge and pay less for doing so. Expert curators generate more value in corrections than they consume in inference, and at the top of that tier the relationship reverses into a revenue share.

Revenue: tiered subscriptions in which active expertise reduces cost, bring-your-own-model integration, premium analytics, and revenue sharing with curators whose contributions carry measurable weight.

4. White label implementations

Branded instances for breed registries and associations, veterinary hospital networks, large livestock operations, and agricultural universities. Custom development, data migration, integrations and dedicated support on the same framework. Each implementation becomes the case study that sells the next.

Development runs in four phases, and the phase boundaries are drawn by what is actually working rather than by what is funded.

The four phases, and what is built today →

See the working AI layer without leaving this summary:

AI & Knowledge Management →

Where to go next

The business case is the near view. If you want the argument for why a knowledge platform for animals matters beyond the next five years, that is a different page and a longer horizon.

Where this goes beyond animals →