A construction-company partner I was on a call with last week put it well: "I get what you do, kind of — but you're not an MSP, you're not a development firm, and you're not a software product. So what are you?" Fair question. The category we operate in doesn't really have a name yet, so let me describe it the way I'd describe it to him: in plain language, with concrete deliverables.
Managed Software Ops is the recurring discipline of running custom software a business depends on. It's a service. You buy it monthly. We deliver it across six standing operational disciplines, applied to your specific system on a continuous basis.
Here's what's actually inside it.
What you actually buy
Monitoring and detection. We watch your system continuously — is it up, are users hitting errors, is anything backing up internally, are the backups actually working, is anything costing more than it should. Anomalies get flagged before you see them. Trends get caught before they become incidents.
Incident response. One number, real people, business-hours response by default. Off-hours coverage is available on the larger tiers. Every incident gets logged, root-caused, and written up. Recurring patterns get pushed back into the system so they stop happening.
Maintenance. Bug fixes, dependency updates, security patches, restore drills, the small stuff. Trivial changes — copy edits, field additions, small UI tweaks, minor logic adjustments — are handled inside the standing fee. If an agent can just do it, it's done. Larger changes are scoped and quoted separately.
Release discipline. A separate test environment, change reviews, a way to roll changes back if something breaks, release notes in plain English, an auditable record of every change that has shipped. No more "I think we deployed something Tuesday."
Documentation. Runbooks, architecture diagrams, data maps, user guides — kept current against the actual code by software agents that watch the codebase and update the corresponding docs whenever something changes. Readable by a non-engineer. Available to you any time. Not in someone's head.
Reporting. Quarterly value review. Uptime, incidents prevented, fixes shipped, spend, risks we're watching, recommendations for next quarter. Thirty minutes, written summary, no theater.
That's the service. Six disciplines, continuous, the same shape on every account we run.
How it's priced
Six fixed monthly tiers, ranging from $500/mo for a simple single-integration system up to $9,500+/mo for a multi-system mission-critical platform. Tier is determined by criticality, integration count, data complexity, and operational demand — not by hours and not by line count. A High-Availability rider is available on the larger tiers for 24/7 on-call and faster recovery commitments.
Enhancements are deliberately separate. They run on capacity blocks, enhancement sprints, or embedded fractional capacity. Support keeps the system healthy. Enhancements move it forward. Different jobs, different budgets, both transparent.
Why this price point is even possible
This is the part that's actually new. Operating discipline at this depth used to require a real engineering team — meaning real engineering payroll. For a 25-person company running one critical custom system, the math never worked. They'd skip it, and just hope.
What changed is that most of the standing operational work — the watching, triaging, classifying, documenting, patching, fixing — is now done by AI agents. Software agents look at every incoming issue, figure out what it is, and route it. They keep the documentation in sync with the code. They handle the routine, low-risk fixes that don't require a human to decide anything. The agents do the volume; our operators handle the things that genuinely require human judgment.
That isn't AI as marketing decoration on a service business. It's a different cost structure. The framework that makes it possible is described separately. The relevant point here is that without the agents, the price wouldn't exist. With them, you get the same level of operating discipline a serious technology company would have, sized and priced for the business that actually needs it.
What you don't buy
You don't buy net-new features inside the support fee. You don't buy redesigns. You don't buy unlimited changes. You don't buy enterprise-style service guarantees at small-business pricing. And you don't buy execution labor that you direct hour-by-hour.
What you buy is an outcome — the system is up, the fixes ship, the documentation reflects reality, the backups have been restored from in the last quarter, the incidents that should have happened didn't. Delivered for a fixed monthly fee, by a team that's contractually accountable for it.
That's Managed Software Ops. Real category, real deliverables, real price. The name is new. The work isn't.