aijakejbPortfolio
Meet Bob, the agent running Beyond IT Support
Beyond IT Support is a managed service provider — the kind of business that runs on a constant stream of tickets, alerts, and small operational decisions. Bob is the agent we built to sit inside that stream, running in production since 2026.
The problem
MSPs live on repetitive, high-volume operational work: triaging tickets, watching monitoring alerts, keeping documentation in sync with what's actually deployed, handling routine account changes. None of it is hard, individually. All of it competes for the same few hours a day, and it's exactly the kind of work that quietly slips when a team is stretched — not because anyone drops it on purpose, but because there's always something louder.
Generic AI chat tools don't fix this. They answer a question when you ask it; they don't sit in the ticket queue and do the work.
What we built
Bob lives inside Microsoft Teams, where the team already works. It reads across the tools an MSP actually runs on — the ticketing system, the RMM/monitoring platform, the documentation store, identity and mailbox systems — and holds a memory of what it has seen and done, so its decisions build on context instead of starting fresh every time.
Every action that changes something — closing a ticket, sending a message, modifying a configuration — passes through an approval gate. Bob proposes; a person decides. What Bob is allowed to do autonomously is a setting, not a fixed line, and it can be tightened or loosened at any time.
What it does today
Ticket triage
Reads incoming support tickets against live monitoring data before a human opens them, so the person who picks it up already has the context instead of starting from a blank ticket.
Alert noise reduction
Watches infrastructure alerts around the clock and closes the ones that resolve on their own, instead of leaving them to pile up as noise a technician has to wade through.
Documentation & config hygiene
Cross-checks what's documented against what's actually deployed, and flags — or drafts a fix for — the gap, rather than letting docs quietly drift from reality.
Communications, held for approval
Drafts client-facing updates and internal notes. Nothing goes out without a person approving it first — the agent proposes, it doesn't publish.
Lifecycle changes
Handles the repeatable steps in account and access lifecycle changes — renames, handovers, offboarding — the parts that are procedural but easy to get subtly wrong by hand.
Daily reporting
Summarizes what it did, in plain English, every day — what moved, what's stuck, and what still needs a decision.
“We didn't start building agentic AI because it was fashionable. We built it to run our own company first — and we still run on it every day.”
Why this matters for the build
Everything on this page is describing software running against a real, operating business — not a demo environment. That's the difference in how Beyond IT Support approaches an agent build: it has to survive real edge cases, real failure modes, and real consequences when something goes wrong, because that's the standard it was held to before it was ever offered to anyone else.
Want to see where this fits in your business?
We start by finding where the repetitive work actually lives — not by pitching a generic rollout.