A Private AI Assistant for Canadian Business
Your team already has the answers written down somewhere. They sit in SOPs, policy PDFs, shared drives, old email threads, and in the heads of the people who have been there longest. The problem is finding them. New hires ask the same questions for weeks. Experienced staff lose time digging for a document they know exists. And when someone reaches for a public AI chatbot to speed things up, your internal information leaves the building and lands on a server you do not control.
A private, self-hosted AI assistant for Canadian business fixes both sides of that. We build an assistant trained on your own documents and standard operating procedures, so staff can ask a plain question and get an answer drawn from your actual material, not the open internet. It runs on infrastructure you control, with role-based access so people only see what they are allowed to see. Every interaction is logged, which matters when an auditor or a cyber-insurance form asks how you handle information.
This is not a public model with your logo on it. Your data stays yours. Nothing is used to train someone else's product, and you decide where it runs.
What is included
What you get with AI Solutions
Assistant trained on your documents and SOPs
We connect the assistant to your own material, including procedures, policies, handbooks, and reference documents, so answers come from your organization rather than the open internet.
Internal knowledge search
Staff ask a question in plain language and get an answer pulled from your documents, with the source it came from so they can check the original.
Role-based access control
Access is scoped by role, so a front-desk user, a manager, and a finance lead each see answers drawn only from the material they are permitted to access.
Onboarding and SOP guidance
New hires can ask how a process works and get walked through the relevant procedure, which cuts the repeated questions that pull senior staff away from their own work.
Audit logs
Every question and answer is recorded, giving you a clear trail of what was asked, when, and by whom for internal review and audits.
Self-hosted deployment
The assistant runs on infrastructure you control, built with self-hosted large language models so your documents and conversations do not leave your environment.
Compliance-ready setup
We map the assistant's access controls and logging to PIPEDA and PHIPA, and align the surrounding security work to the CIS Controls and the NIST Cybersecurity Framework, with an evidence pack you can hand to auditors or insurers.
Why open source
Why we build this on open source
Open source is what makes a private assistant practical for a small or mid-sized organization. Because we build on self-hosted large language models with frameworks like Genkit and LangChain, there are no per-seat licence fees stacking up as your team grows, and you are not locked into one vendor's pricing or roadmap. More importantly, self-hosting is what lets your data stay yours. The model runs on infrastructure you control, so your documents and the questions your staff ask are not sent to a third-party service or used to train someone else's product. We are honest that the real cost is in the build and ongoing support, not a licence, and that is where our work is.
How it works
Getting started
Free assessment
We start with a no-obligation assessment of what your team asks most, where your documents live, and who should be allowed to see what, then scope a quote to fit.
Connect your documents and set roles
We load your SOPs, policies, and reference material into the assistant and define role-based access so each person sees only what they are permitted to.
Deploy on infrastructure you control
We stand up the assistant on self-hosted models, turn on audit logging, and test answers against real questions from your team before going live.
Train your team and support it
We show staff how to use it, hand over the evidence pack for compliance, and offer ongoing managed support at a predictable monthly fee.
FAQ
Questions about AI Solutions
Does our data leave our environment or get used to train other AI models?
No. The assistant is built on self-hosted large language models running on infrastructure you control. Your documents and the questions your staff ask stay in your environment and are not sent to a third-party service or used to train anyone else's product.
How does the assistant keep sensitive information from the wrong people?
Access is role-based. Each user only sees answers drawn from the material their role permits, and every question and answer is recorded in audit logs so you have a record of who asked what and when.
Can this help with PIPEDA or PHIPA obligations?
We map the assistant's access controls and logging to PIPEDA and PHIPA, and align the surrounding security work to the CIS Controls and the NIST Cybersecurity Framework. We also produce an evidence pack of mapped controls, documented policies, and monitoring records for audits and cyber-insurance questionnaires. These are frameworks we align to, not certifications we hold on your behalf.
What does it cost?
We start with a free, no-obligation assessment, then scope a quote to your situation. Because we use self-hosted, open-source models, there are no per-seat licence fees. The real cost sits in the build and ongoing support, and ongoing managed support is a predictable monthly fee.