The frontier AI for IT work
Case studies and field notes from MSPs putting AI on the service desk — what actually closed, what stayed human, and what we learned running it in production.
A production case study, plus three documents written for the people who have to make the decision rather than for a search engine.
934 non-alert tickets in April (611 hours). Modeled coverage expands from 25% today to 66% near-term, unlocking 411 hours of technician capacity per month.
The first public classification of what MSPs actually do. 17 domains, 142 work types, and the rules that keep a ticket in one class instead of another. If you have ever tried to report on ticket categories and found the categories moved under you, this is the document for that problem.
AI as a helping hand writes nicer notes. AI as the service desk closes tickets, up-levels technicians and shows up on the P&L. This brief separates the two and works through what has to be true for the second one to happen in your shop.
Identify the right work, turn it into agent workflows, govern behaviour and data, monitor with human oversight, then run it on the desk. A sequence you can follow rather than a set of principles you can agree with and not act on.
Book 30 minutes. Bring the taxonomy, bring your own ticket categories, and we'll go through where this would and wouldn't hold in your shop.