Pricing
AI Agent Pricing: Fixed Prices, Agreed Before Work Starts
Every agent we build has one fixed price, in writing, before we start, and you see a free working prototype before you pay anything. Here are our starting prices for Canadian businesses, what each one includes, and the running costs you should plan for.
What are your prices?
Starting prices for the production build. Your scope sets the exact figure after the free audit and prototype.
Before you pay anything
A free working prototype in 5 business days
"I'll build you a working prototype in 5 business days, free, no obligation."
Scott Holmes, founder
- When: after the free 30-minute audit, once we agree on the one job to prototype.
- What: one workflow or one agent, on your real tools in a test setup or on sample data. It never touches live customers or production data.
- If you stop there: you owe nothing and keep a written summary of what it does and what the full build would cost.
- If you go ahead: you pay the fixed production price below, and the prototype isn't charged on top. Limited to 4 a month.
One workflow
Single automation
One-time, fixed before work starts
- One job between your apps, like a web form into your CRM or closed deals into QuickBooks
- Built on n8n, Make.com or Zapier, with an AI step where it helps
- Tested before it goes live
- Short documentation
Most businesses start here
First AI agent
One-time, fixed before work starts
- Free audit and a free working prototype first
- Written scope and one fixed price
- An operations, customer or private agent, connected to your tools
- Human approvals and a full activity log
- Staging review before go-live
- Typically live in about 2 weeks
- Connected to one business system; each extra system is from $499
- Documentation, a walkthrough and 2 weeks of support
Optional
Monthly support
Ongoing, after the included 2 weeks
- Monitoring and regular log review
- Fixes when a connected app changes
- Prompt and rule tuning as your business changes
- Small changes and additions
- Hosting and updates for agents we run for you
- Includes 150 AI voice minutes a month plus normal text-model use, with any overage billed at cost
- Month to month: cancel anytime and keep everything we built
What makes one agent cost more than another?
Six things move the build price. We talk through each on the audit call.
🔌 Integrations
How many systems the agent touches, and whether they have a usable API. Reading email and writing to QuickBooks is simpler than touching your ERP, your CRM and a supplier portal in one run.
📄 How messy the input is
Clean, consistent emails are quick. Scanned forms, a dozen supplier invoice layouts or handwritten notes take more testing.
📈 Volume
Ten items a day and five hundred a day need different error handling, monitoring and testing, and they change your model usage.
🔀 Rules and exceptions
Every "except when" in your process is another branch to build and test, and sometimes another approval step.
🛡️ Where the model runs
A hosted model is the quickest to set up. A private model on a Canadian cloud server or your own hardware adds setup and hardening work.
📞 Voice
Phone agents need call-flow design, a phone number, transfer rules and a lot of test calls.
What will it cost to run each month?
Running costs are separate from the build price and depend on your volume. Monthly support includes 150 AI voice minutes a month plus normal text-model use, with any overage billed at cost. We estimate the rest in the scope from your own numbers.
- Model usage: hosted models such as Claude and OpenAI charge for the text the agent reads and writes. A long PDF costs more to process than a short email, and a larger model costs more per request than a smaller one.
- Voice minutes and phone numbers: Retell AI and Vapi charge per minute of call time, and Twilio charges for the phone number and texts.
- Workflow platform: a subscription to n8n, Make.com or Zapier, or a small server if n8n is self-hosted.
- Private hosting: for private agents, a Canadian cloud server billed by the hour or month, or your own hardware and power.
- Monthly support: optional, from $99/month.
Two ways to handle usage. With monthly support, we run the usage-based services and include 150 AI voice minutes a month plus normal text-model use, with any overage billed at cost. If you would rather hold the accounts yourself, we set them up in your business's name, the bills come to you directly, and the support price stays the same.
What's included, and what isn't?
Included in every agent project
- Audit: a free 30-minute call by phone or video.
- Prototype: a free working prototype in 5 business days, before you commit.
- Scope: a written scope with one fixed price.
- Build: connections, instructions, approvals and logging.
- Staging: a review copy you test before go-live.
- Handover: documentation and a walkthrough for your team.
- Support: 2 weeks after launch.
Not included
- Usage beyond the allowance: model usage, voice minutes and software fees beyond what monthly support includes, or billed to you directly if you hold the accounts yourself.
- Hardware: servers or GPUs for private agents.
- Work outside the scope: quoted separately, at a fixed price, before we do it.
- Data clean-up: fixing years of messy records beyond what the agent needs to work.
Pricing questions
Why don't you charge by the hour?
Because you should know the cost before you commit. We scope the job first, then give one fixed price. If the build takes us longer than planned, that's our problem, not yours.
Are model usage fees included in the price?
With monthly support, yes, up to an allowance: 150 AI voice minutes a month plus normal text-model use, with any overage billed at cost. Without monthly support, model usage is billed by the provider, such as Anthropic or OpenAI, based on how much the agent reads and writes. Either way, we estimate the monthly amount from your volume in the scope, and the activity log shows what the agent used.
Do we have to buy monthly support?
No. Every project includes documentation, a walkthrough and 2 weeks of support after launch. Monthly support is optional, for businesses that want us to monitor the agent, keep it working when connected apps change, and make small improvements. It's month to month, and you can cancel anytime.
If we go ahead after the free prototype, do we pay for it?
No. If you go ahead, we build it into production at the fixed price on this page, and the prototype work is not charged on top. If you don't go ahead, you owe nothing and keep a written summary of what the prototype does and what the full build would cost. The working prototype itself stays with us.
Is the agent audit really free?
Yes. The 30-minute audit is free, and you keep the list of recommended jobs whether you hire us or not.
What if the scope changes partway through?
We tell you before doing anything outside the agreed scope, and you get a separate fixed price for the change. You decide whether to add it now, later or not at all.
How do our prices compare? We checked what other Canadian agencies publish for custom AI agents (October 2026), with links to every source. See the price research.
Want a fixed price for your first agent?
Book a free 30-minute agent audit. If there's a job worth automating, I'll build you a working prototype in 5 business days, free, and you get one fixed price in writing for the production build.
Get a Free Working Prototype