Managed services

Giving a service desk AI over the systems it already runs

A managed-services firm ran its customers from six systems that did not talk to each other. Every question meant six logins. Instead of an app, we put AI tooling in the technicians' hands that reaches all six, under a written policy for what it may read, and one rule: every change is a plan a person approves before it runs.

Systems reached

six

Changes without a person's yes

none

Apps built

none

What was wrong

A managed-services firm looked after dozens of customers from six systems that did not talk to each other: tickets in one, device monitoring in another, security alerts in two more, the customers' cloud tenants, the network. A simple question, which of this customer's machines are missing a security agent, meant six logins and a spreadsheet. Staff were already trying AI tools on their own, with no rule about what could be pasted where.

What we did

No new app. We gave the technicians AI tooling that reaches each of the six systems directly, resolves a customer by name across them, and answers in plain terms with the next step included. Reads are open. Every write is a plan the tool shows, a person approves, and the tool then applies, with a receipt. A written policy says which data the AI may reach, where it runs, and what stays out; secrets stay where the firm keeps them.

What it got them

One question, one answer, across six systems, in use on the lead's seat today, with the desk-wide rollout next. The policy came before the tools, so the firm knows exactly what its AI touches.

You do not always need an app. Sometimes you need the right tool, set up properly, with a policy.

Anonymized. A managed-services firm; name withheld, its systems named by kind only. Engagement led by Ryan Luccioni, principal.

← All work