A system nobody watches quietly stops working
We monitor your systems, fix what breaks, and keep improving them. Monday to Friday, 9am–6pm MT, with response commitments in writing rather than best effort.
Response commitments
Improvement hours included
From $750/mo
A support and maintenance plan keeps deployed AI systems, automations, and applications running: uptime and error monitoring with alerting, dependency and model updates as providers change them, a committed response time when something breaks, and a monthly allocation of improvement hours. DAvision plans run from $750 per month for monitoring with a two-business-day response to $3,000 per month with a four-hour response, twelve improvement hours, and written post-mortems.
AI systems decay faster than software used to
Model providers change things underneath you
An API version is deprecated, a model is retired, a rate limit changes. Nothing on your side moved, and yet it broke.
Costs drift without anyone watching
Token and API spend creeps upward. Most teams find out from the invoice, a month late.
The person who built it left
Undocumented systems become untouchable. Everyone works around them rather than fixing them.
What a plan covers
Uptime and error monitoring with alerts
We watch it, and we find out before you do.
Committed response times
Two business days, eight business hours, or four business hours depending on the plan. In writing, not aspiration.
Model, library, and dependency updates
When a provider deprecates something we migrate it before the deadline rather than after the outage.
Improvement hours every month
Two, five, or twelve hours of actual development, so the system keeps getting better rather than just staying alive.
Cost monitoring
Cloud and API spend tracked and reported, with alerts when it moves unexpectedly.
Post-mortems after incidents
At Enterprise, a written account of what happened, why, and what changed so it does not recur.
How it works
Four steps, no surprises. You see a working version before you pay the balance.
Onboarding review
We document what exists, what it depends on, and what failure looks like for each part.
Instrument and alert
Monitoring and alerting set up on every critical path.
Run the plan
Monitoring, updates, and improvement hours delivered monthly with a health report.
Review regularly
Quarterly on Pro and Enterprise: what broke, what improved, what to do next.
Packages & pricing
Fixed-price builds in Canadian dollars. Ongoing support is quoted separately from $750 per month.
- Uptime and error monitoring with alerts
- Response within 2 business days
- Model, library, and dependency updates
- 2 improvement hours per month
- Monthly health report
- Everything in Basic
- Response within 8 business hours
- 5 improvement hours per month
- Cloud and API cost monitoring
- Quarterly review call
- Everything in Pro
- Response within 4 business hours
- 12 improvement hours per month
- After-hours emergency line, billed hourly
- Written post-mortem after every incident
- Direct line to your architect
AI systems decay differently from ordinary software
Conventional software that nobody touches tends to keep working. AI systems do not, because the ground moves underneath them. Model versions get retired on a published schedule and sometimes ahead of it. API endpoints are deprecated. Rate limits and pricing change. A system that was correct and affordable in March can be broken and expensive by September without a single line of your code changing.
The second decay path is cost. Token and API spend drifts upward as usage grows and as prompts accumulate context nobody pruned. Most teams discover this from an invoice rather than a dashboard, a month after it started. Cost monitoring at Pro tier tracks spend against a baseline and alerts on unexpected movement, which usually pays for a meaningful part of the plan on its own.
The third is organisational. The person who built the automation leaves, nothing was documented, and the system becomes untouchable — everyone works around it instead of fixing it, and the workarounds accumulate until the automation is producing negative value. Onboarding starts with documenting what exists and instrumenting it, which is the antidote to exactly this.
The improvement hours matter more than they look. A support plan that only fixes breakage leaves a system frozen at the capability it launched with, while the tooling underneath it improves every quarter. Two to twelve hours of actual development each month means the system compounds rather than stagnates — which over a year is usually the larger share of the value.
What changes after launch
Concrete, measurable, and checkable against the baseline we take before we start.
Provider changes handled before the deadline
Deprecations get migrated ahead of the cutoff rather than after the outage.
Spend that stays visible
Cloud and API costs tracked against a baseline, with alerts when something moves.
A system that keeps improving
Monthly development hours mean capability grows instead of freezing at launch.
Who this is for
Anyone running an AI system in production. Businesses whose automation has become load-bearing. Teams without an in-house engineer who can respond when an API changes.
We work with dental clinics, construction, energy, real estate, and commercial businesses across Calgary and Canada.
Frequently asked questions
What does a support plan cost?
Monitoring, alerting, dependency updates, and two improvement hours per month start at $750. Eight-business-hour response with five improvement hours and cost monitoring is $1,500 per month. Four-business-hour response with twelve improvement hours, an after-hours line, and post-mortems is $3,000 per month.
Do you support systems you did not build?
Yes, after an onboarding review to document what exists and instrument it. If the system needs hardening first, we say so and quote that separately rather than absorbing it into a monthly fee.
What are the support hours?
Monday to Friday, 9am to 6pm Mountain Time. Enterprise includes an after-hours emergency line billed hourly.
What counts as an improvement hour?
Actual development — a new workflow step, a tuned prompt, a report change, an added integration field. Not monitoring or routine updates, which are included separately.
Can we change plans?
Monthly, in either direction, with notice before the next billing cycle.
What if nothing breaks?
You still get the improvement hours, the updates, and the health report. The plan is maintenance and ongoing development, not an insurance policy that only pays out on failure.
Related services
Thirty minutes, no pitch
We map out exactly what you need and what it would cost. If automation is not the answer, we will say so.
