Why ACS
We built it before we sold it.
ACS Automate — the platform we run on ourselves
We operate a managed services business: tickets, alerts, devices, clients. Two years ago, we built an AI operations platform — internally called ACS Automate — to run our own service desk better. Today it is a stack of production AI workloads we depend on daily. It informs every part of the MIP practice.
A short, partial list of what ACS Automate does for ACS:
- Autotask ticket enrichment. Every inbound ticket gets a structured triage — category, priority recommendation, suggested owner, recent history with this client and device — within seconds of arrival. Roughly 1,200 tickets per week pass through enrichment. We measure the override rate (how often our tech changes the recommendation) and tune the model against it.
- Autonomous remediation. Categories of Datto RMM alerts — disk usage, Windows Update failures, AV/EDR posture drift, certain printer and service-restart conditions — fire approved remediation playbooks automatically when device and risk criteria match. Workstations auto-approve at our threshold; servers require human approval. The pipeline runs in production across 23+ client environments.
- Microsoft Teams bot for the service desk. Our techs create tickets, look up devices, search the ITIL knowledge base, and triage alerts inside Teams. The bot uses Microsoft App Registration auth, talks to our internal Postgres + Vault + Cloudflare Access stack, and posts daily Minimum Viable Quality (MVQ) close-coaching DMs to individual technicians based on ticket scoring.
- ITIL knowledge base with retrieval. 36 internal docs ingested into a structured store, queryable from the Teams bot. When a tech asks “what’s our process for X,” the bot answers from the knowledge base, not the model’s general training.
- Quarterly Technology Business Reviews. Our TBR Report Generator pulls from Autotask, Datto, ConnectSecure, IT Glue, and Pax8 to produce a 60-section quarterly client review. Async job pipeline, orphan-reaper, written narrative. We use it on our own clients. We don’t outsource the deliverable.
- Enrichment backfill. A nightly process that catches tickets dropped by webhook glitches during high-volume creation bursts. Last month it recovered six missed tickets that would otherwise have shipped without enrichment.
This is not a sales claim. This is the operational reality of our business. The MIP practice exists because we built and ran this platform first. The Custom Agent Sprint methodology, the workflow instrumentation, the human-in-the-loop pattern, the model-version-change discipline — all of it came out of running ACS Automate in production. We are selling the discipline, not a product demo.
Michigan-rooted, remote-fluent
Headquartered in Traverse City, Michigan. Service region: Northern Michigan and Central Florida for on-site work; remote across the country for everything else. Real humans answer the phone at 231-267-TECH (8324) during business hours.
We are not a Silicon Valley vendor pretending to understand small business. The partners and operators in our client base run real companies and have no patience for AI vendors who don’t. We share the patience. We share the conservatism. We share the operating-cost discipline.
The team
Not bios. What the team ships.
The MIP practice is staffed by the same engineers who maintain ACS Automate, who write our remediation playbooks, who handle the Tier 2/3 escalations on our MSP side, and who have produced quarterly Technology Business Reviews for our managed services clients for years. AI is not a side hustle for us. It is the operating system of our own business, and the discipline that runs our own service delivery. When you become an MIP client, you get the same engineering team and the same operating discipline.
The named lead on your account is not a sales engineer. It is the person whose work will show up in your quarterly ABR.
How we’re different from your other options
Honest comparison. No logos, no specific competitor names.
MSPs adding AI to their stack. Most MSPs resell Microsoft Copilot and add a slide to their quarterly review. That’s not a practice; that’s a feature add. There are good ones who are building real AI capability, but they’re rare today and you’ll be able to tell within 30 minutes of a discovery call. We were one of them, then we built ACS Automate, and then we stood up MIP as a separate practice. The two are linked, but the practice is its own discipline.
AI consultancies. Strong at engagement, weak at operation. The typical engagement is project-shaped — a build, a handoff, a goodbye — and the AI program drifts after they leave. Some consultancies are excellent for the build phase; most leave you operating it. MIP is built around the operate phase, and we expect to be running it three years from now.
Big-firm AI advisory. Designed for enterprise. The methodology is usually right and the price is usually wrong for SMBs. If you’re 500+ users with a CIO and an internal data team, big-firm advisory might be a fit. If you’re 25–200 users with a managing partner and an IT person, it isn’t.
DIY. Possible. We have clients who have a CTO with bandwidth and good instincts, and we serve them differently — more on the governance and measurement side, less on the build side. Most SMBs don’t have the bandwidth. If you’re one of them, MIP makes the time problem go away. If you have a strong internal lead, we’ll tell you and scope accordingly.
Book an AI Readiness Assessment
Ready when you are.
Three weeks. Fixed fee. A real report and a 90-day roadmap. You keep it whether you hire us or not.