What is a digital worker?
A digital worker is an AI agent that does real work in your business, not just answers questions in a chat window. It reads your inbox, looks at your calendar, queries your CRM, drafts replies, schedules follow-ups, and writes to the systems you use every day. Where a chatbot answers a question, a digital worker completes the task that question was about.
How much does an AI agent cost for a small business?
Our tool consultation is $750 for a half-day, single-tool deep dive where we sit down and walk through one tool you already use and how to get more out of it with AI. Our install tier is $1,500 for a one-day installation plus $1,800/month for the running infrastructure. Mid-tier engagements where we spend a week on-site connecting every tool you use are $10,000-$12,000 one-time, no monthly fee after, because the point is we teach you the plumbing and you own it. Sovereign local-AI deployments where we build out the hardware so you can run your own local models inside your building run $20,000 and up depending on capacity. No surprise bills, no per-seat licensing, no per-message charges.
Can your AI agents work with QuickBooks, HighLevel, Stripe, or other tools we already use?
Yes. The AI does not replace your existing tools, it works with them through their APIs. We have working integrations with QuickBooks Online, Stripe, HighLevel, Google Workspace, Microsoft 365, Twilio, and most CRMs that have an API. Meta Ads, Google Ads, and other ad-platform integrations are on our roadmap. If a tool you use does not have an API, we still find a way in most cases.
Is OAL based in Canada?
Yes. Obsidian AI Labs Inc. is incorporated in Ontario, Canada, based in St. Thomas. We serve customers across Canada and remote-friendly customers globally. Pricing is in Canadian dollars.
How is OAL different from hiring a developer or signing up for AI SaaS?
A developer builds you software. AI SaaS sells you a feature behind a login. OAL plumbs together what you already pay for, your AI subscriptions, your CRM, your books, your email, into a system that does the work between your decisions. You pay once for the plumbing, then a small monthly cost for the infrastructure that keeps it running. You own the configuration. Nothing is locked behind another vendor's app.
Do I need to know how to code to use OAL?
No. You interact with the AI through normal interfaces, Telegram, email, a web dashboard, a phone call. The code lives on the agent's side. You ask a question in plain English, the agent does the work. If you want to read and tweak the underlying configuration, we make that possible too, but we never require it.
How long does an AI implementation take?
A Tier 1 install is one day. A Tier 2 concierge week is five working days. A Tier 3 sovereign deployment is roughly two weeks of OAL labor plus the time it takes hardware to arrive, one to three weeks depending on what we order. Most customers see real work happening from the agent inside week one.
What happens to our data?
It stays where it already lives. The agent talks to your existing tools through their APIs, it reads your CRM record from your CRM, your email from your email provider. Nothing duplicates into a third-party data warehouse OAL owns. For customers with compliance constraints, two options: our Tier 3 sovereign deployment runs the AI itself on hardware in your building so even model inference never leaves the network, or for customers who need a high-end hosted model but cannot use the public Anthropic or OpenAI endpoints, we deploy through Amazon Bedrock which handles SOC compliance, BAA, HIPAA, and most regulated-industry requirements while still using the same class of model.
What if the AI makes a mistake?
Every action the agent takes is logged. If the agent drafts an email, sends a calendar invite, or writes to your books, you can audit exactly what it did and why. We default to letting the agent draft but require human approval before anything is sent or charged. As trust builds for repetitive tasks, you can lift that gate per workflow. The point of the system is to make AI accountable in a small business the same way an employee is.
Who is Brandon Dietz?
Brandon is the founder of Obsidian AI Labs. A decade in cybersecurity, currently Manager of External Network Security Services at Digital Boundary Group, Fanshawe College graduate, based in St. Thomas, Ontario. He started OAL after looking at the AI tools small businesses were being sold and realizing none of them actually did the work, they only answered questions about the work. OAL exists to fix that.
Can I see what OAL has actually built before we talk?
Yes. The OAL blog at obsidianailabs.ca/blog walks through real work done by the agents. A CFO agent that connected QuickBooks for the first time. An agent that traced an unfamiliar charge back to its own toolbox in thirty seconds. An agent that pulled doorbell footage and identified a door-to-door salesperson in sixty seconds. The blog is the receipt that this is real work, not a marketing demo.
What kinds of businesses does OAL work best with?
Owner-operated small businesses where the founder is doing too much admin themselves. Professional service firms where high-skill people get pulled into low-skill follow-up work. Non-profits that need to compete with bigger orgs without the overhead. If you can describe a repetitive task that eats your week and you wish a colleague could do, that is the shape of work OAL automates.
Are you available to do this for businesses outside Ontario?
Yes. We travel to customers. On-site is our preferred way to install and configure because the wins compound when we're in the room together watching how your team actually works. Some remote options are possible for follow-on adjustments and the tool consultation tier. Tier 2 concierge weeks and Tier 3 sovereign deployments are always on-site. We cover all of Canada, and case-by-case for US-based customers.