How we work
No generic tools. No data exposure. A deliberate process that ends with AI that actually knows your business — and stays that way.
Book a 30-Min Private AI Discovery Call →We've built AI deployments for law firms, manufacturers, and healthcare practices. Every one starts the same way — with us listening before we write a line of code.
A 30-minute call to understand your business, pain points, and what you want AI to actually do. We tell you plainly whether AI is the right tool and what it will cost. No pressure, no pitch deck.
30 min · FreeWe audit the data you have — documents, databases, emails, SOPs, CRM exports — and tell you exactly what we can train on and what gaps need filling. You get a written assessment before any contract is signed.
1–2 weeks · Included in fixed-fee tierWe build and fine-tune your AI on your data, in a private environment we control. Your data never touches a public LLM. We run it by you with real prompts before you commit.
3–8 weeks depending on scopeWe deploy to your environment — on-premise server, private cloud, or secure VPC — and run hands-on training with every employee who will use it. We don't hand you a user manual and disappear.
1–2 weeks · Included in all engagementsYour business changes. Your AI should too. Retainer clients get quarterly model refreshes, usage reviews, and direct access to our team. We track output metrics so you can see exactly what you're getting.
Monthly retainer · OptionalTwenty years of Vermont IT services taught us one thing: businesses want to know what they're getting and what it costs. These are our three packages.
One-time · Delivered in 2 weeks
Best for: Businesses not yet sure where to start. Applies as a credit toward any build engagement.
Start with an Assessment →Project-based · Scoped after assessment
Best for: Businesses ready to deploy AI that actually knows their operations.
Get a Project Quote →Monthly · Min 6-month term
Best for: Post-build clients who want their AI to stay current and their team supported.
Talk About Ongoing Support →This is the question every business should be asking before they use any AI tool. Here's exactly how we handle your data — in plain language, not legal boilerplate.
Your AI runs on infrastructure you control — either on your own servers, a private cloud environment we configure and hand you keys to, or a dedicated private VPC. It is never co-tenanted with other clients. We design for the assumption that the network perimeter will be breached; your data stays safe anyway.
During the build, your VAS engineering team has scoped access to complete the project. Access is logged and revoked at project close. After deployment, access is entirely yours. We can provision ongoing support access — time-limited, logged, and revocable — only if you explicitly grant it under the retainer.
We start from a base open-weight model (no OpenAI, no Anthropic, no Google APIs in your data pipeline), then fine-tune it on your documents, records, and knowledge base. Your data is used only to train your model. It does not improve any public model. It does not leave your environment. It does not go to any third-party API.
You own it. The trained model weights, the fine-tuning dataset, the deployment configuration — all of it transfers to you at project close. We retain no copies. If you cancel the retainer, you keep running your own AI. There is no vendor lock-in because we built it to be yours from day one.
Everything we build for you is work-for-hire. You own the model, the agents, the integrations. Our engagement agreement spells this out in plain English — no buried IP assignment clauses, no licensing dependencies. If it runs in your environment, it's yours.
Our infrastructure designs are built with HIPAA, SOC 2 Type II, and state-level data privacy regulations in mind. We're not lawyers and we're not your compliance officer — but we are experienced enough to design systems that make your compliance team's job easier, not harder. We document every data flow.
"Trained exclusively on your data" means your competitors, your cloud vendor, and OpenAI will never see what you've built. That's the whole point.
Ask Us a Specific Privacy Question →Private AI matters most when the data is sensitive, the stakes are high, and generic tools aren't cutting it. These are the verticals we know well.
AI trained on your case files, briefs, and precedents — so associates draft faster and partners review less. Client privilege stays intact.
AI that knows your parts catalog, supplier contracts, and production SOPs — cutting quoting time, reducing errors, and keeping institutional knowledge when employees retire.
HIPAA-ready AI for clinical documentation, patient intake, and internal operations — without any PHI touching a public system.
AI for client reporting, regulatory documentation, and internal research — with the audit trails and access controls your compliance team actually wants to see.
Not in one of these verticals? We've also worked with professional services firms, logistics companies, and multi-location retail. If you have proprietary data and a real process, we can probably help. Ask us first.
Not ready to talk yet? Read the Buyer's Guide →
The discovery call is 30 minutes. We'll tell you whether private AI makes sense for your business, what it would cost, and what the data looks like. No pitch, no commitment. If it's not the right fit, we'll say so.
or email us directly at info@vermontaisystems.com