I install, configure, and monitor AI bots running on your infrastructure — so you get custom AI automation without the infrastructure headache, the DevOps contractor search, or the “nobody on our team can maintain this” trap six months later.
This is Auburn AI’s Managed Bot Setup service. One fixed-price installation + a predictable monthly retainer. Built for Canadian SMBs who want AI bots running in production without becoming infrastructure shops themselves.
What this service actually is
You already want AI doing something specific in your business — triaging inbox, drafting proposals, summarizing meetings, answering website visitor questions, qualifying leads, running a scheduled content workflow. You’ve read the blog posts. You’ve watched the demos. You’ve maybe even spun up an n8n workflow yourself and realized that “ship it to production and trust it” is the hard part.
I ship it to production and monitor it so you can trust it.
For a fixed setup fee I build the exact bot or workflow you need, deploy it on infrastructure we agree to (your existing VPS, a new droplet I’ll provision, or Auburn AI’s managed hosting — your choice), and hand off complete documentation. Then for a flat monthly retainer I watch the logs, fix breakage from upstream API changes, and apply small improvements as you identify them.
No vague “AI strategy deck.” No $30K three-month discovery engagement. One working bot, running in production, for money you can justify against its measurable impact.
Pricing
One-time setup: $1,997 CAD
Covers:
- 90-minute intake call to define the bot’s exact scope, inputs, outputs, and success criteria
- Full implementation in n8n (self-hosted on your server or Auburn AI’s managed instance, your preference)
- Claude or GPT-4 integration with your specific prompt library
- Connection to your existing tools (CRM, inbox, Slack, Asana, Google Drive, whatever you use)
- Error handling, logging, and monitoring dashboards
- Two revision rounds after the first working version based on your feedback
- Handoff documentation your team can actually use
- 30-minute team training video covering maintenance + customization
- 14-day bug-fix warranty after launch
Timeline: 2-3 weeks from deposit to production. 50% deposit, 50% on delivery. Stripe or Interac.
Monthly retainer: $197 CAD per bot per month
Starts after the 14-day warranty ends. Includes:
- Uptime monitoring (alerted within 10 minutes of any failure)
- Monthly review of logs for prompt drift, error spikes, upstream API changes
- Upstream API fixes when Claude/OpenAI/Anthropic/Google change behaviour (happens 2-6x per year)
- Up to 2 hours per month of small tweaks or improvements you request
- Quarterly check-in call to review results and suggest optimizations
- Priority email support (business-day response guaranteed)
Month-to-month. No contract. Cancel any time with 30 days’ notice. If a bot stops earning its keep in your operation, you pause the retainer and keep the code.
Multi-bot discount
Three bots = $497 CAD/month flat (instead of $591). Five bots = $747 CAD/month flat (instead of $985). Setup for additional bots after the first: $1,497 each.
Bots I build most often
Scoped, deployed, and running across multiple Canadian SMB clients. Each of these takes 2-3 weeks to ship and fits the $1,997 setup tier.
Inbox triage + response drafter
Your team’s shared inbox gets triaged automatically — urgent vs non-urgent vs sales vs support vs noise. For each non-routine message, a first-draft response is pre-written in your voice (using a brand-voice document we capture in the intake call). Your team reviews and sends from their normal inbox. Saves 5-15 hours per week per teammate on email.
Proposal and quote auto-drafter
Your client intake form feeds into a Claude-based workflow that produces a first-draft proposal in your template, pricing, and voice. Your partner or account exec reviews and polishes before sending. Drops proposal turnaround from days to the same afternoon. Most valuable for professional services firms (legal, accounting, consulting, construction).
Meeting-to-action-items pipeline
Zoom/Google Meet/Teams transcripts flow into a workflow that extracts action items, owners, and deadlines, then writes them directly into your project tool (Asana, ClickUp, Notion, Monday). Cuts the post-meeting documentation tax to near zero.
Website FAQ and lead qualifier
A Claude-powered chatbot on your website answers common questions trained from your site content, qualifies serious leads with a short conversation, and routes them to your calendar or intake form. Reduces unqualified discovery calls and makes after-hours visits convert instead of bounce.
Weekly KPI and status reports
Pulls data from 2-4 sources (Stripe, your CRM, Google Analytics, project tool), passes it through Claude, and delivers a written weekly summary email to you or your executive team. Replaces the 90-minute Monday exec meeting with a 4-minute read.
Custom workflows
Anything with clearly defined inputs and outputs fits this tier. I’ll tell you honestly during the intake call if your use case is better suited to a different approach.
What this is not
- Not chatbot-as-a-service for end consumers. I build workflow automation for inside your business. Not e-commerce support bots for your customers (different problem).
- Not real-time voice AI. Voice AI is its own specialized space. If you want an AI phone agent, I’ll refer you to specialists.
- Not replacing human judgment in high-stakes decisions. AI drafts, humans decide. If you want AI to autonomously approve client invoices, sign contracts, or commit your business to actions, that is not what I build.
- Not 24/7 emergency support. Monitoring is continuous but response is business-day (9-5 Mountain Time). If you need 24/7 on-call, enterprise tier has that — ask.
The honest part
Very few things in business automation are “100% autonomous.” If an AI provider changes their model behaviour (which they do, sometimes with 24 hours’ notice), prompts that worked perfectly on Tuesday can output garbage on Friday. The retainer exists because those changes need to be caught and fixed. If you tell me you want a bot that will run forever without any ongoing cost or maintenance, I’ll tell you the truth: that’s not realistic in 2026.
What is realistic is 85-95% autonomous — the bot handles its normal work without any intervention on your side, and I handle the occasional 5-15% that needs an eye on it.
Who I work with
- Canadian SMBs with $500K-$10M annual revenue
- Professional services (legal, accounting, consulting, healthcare practice, real estate, construction)
- Owners and partners (not employees evaluating for their manager — too much handoff friction)
- Comfortable with async collaboration (Loom walkthroughs, scheduled video calls, email)
- Willing to contribute brand voice + business context during intake (critical for bot quality)
Process timeline
- Free 30-minute discovery call. We confirm fit or I refer you elsewhere. No sales pressure — if it is not a fit, the fastest way I help you is by telling you so.
- Scope document + 50% deposit. Within 48 hours of deposit I start building.
- Week 1-2: build. First working version in your hands by day 10-14.
- Week 2-3: revisions. Two rounds of feedback-and-adjust.
- Week 3 handoff. Production deployment, team training video, documentation, balance invoice.
- 14-day warranty window. Any bug fixes free of charge.
- Retainer begins (optional). Monthly. Cancel any time.
Frequently asked questions
Who owns the code?
You do. Full source code, documentation, and deployment scripts are yours at handoff. If you cancel the retainer, you keep running the bot on your own infrastructure. I do not retain lock-in.
What if we want to bring this in-house after a while?
Excellent — that is often the right outcome. The bots I build use n8n (open-source) + Claude or GPT APIs (vendor-neutral) specifically so you can transition to an internal developer or cheaper ops contractor without rebuilding anything. I will do a 1-hour handoff call to your new person free of charge.
What about data privacy and PIPEDA?
All workflow architecture accounts for Canadian privacy law. When a bot sends data to US-based AI providers (OpenAI, Anthropic), I will document the data flow, assess the risk category, and recommend appropriate mitigations — enterprise agreements with data-handling guarantees, automatic redaction before AI calls, or keeping PII out of the AI layer entirely. Privacy is built into the bot architecture from day one.
How is this different from just buying a SaaS product?
SaaS products are built for the average customer and priced accordingly. Custom bots are built for your specific workflow, CRM, and brand voice. If your work is non-standard enough that off-the-shelf SaaS does not quite fit — which is most professional services firms — custom is more effective. If off-the-shelf genuinely works for you, you do not need this service. I will tell you during the discovery call.
Can we start with one bot and scale?
Yes — most clients do exactly that. First bot proves the model in 30 days. Second bot joins the retainer at 3-6 months if the first has paid for itself (which it usually does within the first 90 days for knowledge-work firms).
Do you offer a free trial?
No free trial. Instead: the discovery call is free, the scope is fixed before payment, and the 14-day bug-fix warranty gives you concrete protection on the setup fee. If the bot does not work to the spec we agreed on, I fix it without additional charge.
Contact
Email: alexander.mcgregor7@outlook.com
Subject line: “Managed Bot Setup — [brief description of what you want automated]”
Response: within one business day, often same-day in evening (Mountain Time).
Thanks — Alexander
Auburn AI is a sole proprietorship based in Calgary, Alberta. Canadian clients invoiced in CAD. International clients invoiced in USD at current exchange.