Should a ticket AI closes cost the same as one a person closes?
Miles Feinberg
Picture two tickets on the same Monday morning. A user who's locked out of their account gets a reset link from the MSP's AI agent and is back in under a minute without a technician touching it. Down the hall, a payroll workstation won't connect to the finance system, and a technician spends 90 minutes on it. On a per ticket invoice, both of those can show up at the same price.
My answer is no. If an MSP bills by the ticket, the price should reflect who did the work. And once you look at why, you'll find a bigger problem with per ticket pricing than AI.
One price for two very different costs
Per ticket pricing made sense when every ticket meant a person's time. AI changes the cost underneath it. A password reset, a license request or a mailbox rule that an AI agent closes on its own costs the MSP far less than an hour of a technician's time. If the rate stays the same, the savings from automation stay with the MSP, and you're paying technician rates for work software did.
Fix one: two tiers
The simplest fix is two rates: a lower one for tickets AI resolves on its own and a higher one for tickets a technician works. For that to be fair, the MSP has to define "AI resolved" clearly (a ticket AI hands to a person halfway through should count as technician work), report both groups every month, and count a reopened ticket against the original instead of billing it twice. If a finalist can't show you a report that splits the two today, they can't bill that way fairly tomorrow.
Fix two: price on business outcomes
The better fix, in my view, is to stop pricing tickets and price the results you're actually buying. That could be uptime on the systems that run your business, how fast people get back to work after an incident, or fewer repeat problems each quarter. It takes more work to set up, since you and the MSP have to agree on how each measure is tracked. In return, it pays the MSP for the thing you want.
The incentive problem
The biggest issue with per ticket pricing has nothing to do with AI. An MSP that bills per ticket earns more when you have more tickets. You want the opposite: as few problems as possible, and the ones that do happen fixed for good. Those two interests pull in different directions every month.
A good MSP will still try to prevent problems, but the pricing model doesn't reward it. Eliminating a recurring printer issue, or automating it away, shrinks their invoice. Under outcome pricing, that same work improves their margin. When you're comparing finalists, that alignment is worth more than a lower per ticket rate.
Questions to ask every finalist
- Do you bill tickets AI resolves at the same rate as tickets a technician works?
- How do you define an AI resolved ticket, and can you show a redacted monthly report that splits the two?
- Which business outcomes would you price against, and how would you measure them?
- When you eliminate a recurring issue, what happens to our bill?
The answers tell you whether a finalist's pricing gets better for you as their AI improves, or only for them.
Talk it through
If you're weighing per ticket, two tier and outcome based proposals, book a free advisory working session with ITBluPrint. We run vendor-neutral MSP selection and don't sell managed services.
