AI MSP INDEX

    AI-first ticketing: the AI sees it before a human does

    This criterion is about order of operations, not outcomes. Whatever the channel, the first thing that touches the ticket is an agent.


    The criterion

    AI-first ticketing. Channels to lodge a ticket request hit an AI before a human. Whether a client submits by email, portal, or phone, the first point of triage is an AI agent that classifies, enriches, and either resolves or routes the ticket. An engineer enters the workflow only when the AI flags that human hands, words, or relationships are needed.

    What it looks like in practice

    A client can raise a request by email, portal, chat, an integration inside a productivity suite, or by picking up the phone. In an AI-first operation every one of those paths lands on the same agent before it lands on a person. The agent classifies the request, enriches it against the client's environment record and history, determines urgency, and either resolves it or routes it with the context already attached.

    The phone channel is where this criterion separates the operations that mean it from the ones that mean it mostly. Handling email and portal tickets with an agent is now ordinary. Putting voice through the same triage, whether by a voice agent or by transcribing and classifying before the call is dispatched, is a materially harder engineering problem, and it is the channel where the human-first habit survives longest.

    Routing is the part that is easy to underrate. A triage agent that only tags tickets has automated an administrative step. A triage agent that decides whether human hands, human words or a human relationship are required, and routes accordingly, has automated a judgement that used to require a senior technician reading a queue. The second is the criterion. The first is a tagger.

    An engineer entering the workflow is not a failure state. The criterion says the AI goes first, not that the AI goes alone. What it rules out is a service desk where a person reads the inbox, decides what each ticket is, and hands some subset of it to a tool.

    How it differs from AI ticket resolution

    These two criteria are adjacent and routinely conflated. AI ticket resolution is about outcomes: what proportion of work closes without a person. AI-first ticketing is about sequence: what touches the work first.

    An MSP can satisfy either one without the other. A service desk can run every intake channel through an agent that classifies and routes well while closing almost nothing autonomously, because its client base skews toward requests that need hands on hardware or a conversation. Equally, an MSP can auto-resolve a large slice of password and access requests through a self-service portal while its email queue is still read by a person each morning.

    We assess them separately because they fail separately, and because conflating them is how an operation talks itself into believing it is further along than it is.

    How the index assesses it

    We enumerate the intake channels and ask, per channel, what the first system to process the request is. A single channel where a human reads first is not disqualifying on its own, but it needs to be named rather than glossed.

    We look for evidence that triage output is used downstream. Classification that nobody trusts gets silently overridden, and an operation where technicians reclassify most tickets on pickup does not meet this criterion however sophisticated the agent is.

    Useful evidence includes a redacted ticket showing the agent's classification and enrichment at the top of the record, the routing rules the agent's output feeds, and an honest account of the override rate.

    Signals and anti-signals

    SignalAnti-signal
    Email, portal, chat and phone all pass through triage before dispatchEmail and portal are agent-triaged, the phone still rings a technician, and nobody mentions it
    Triage enriches from the environment record and ticket history, not just keywordsClassification is keyword matching relabelled as AI
    Routing decides between hands, words and relationship, then dispatchesThe agent tags the ticket and a dispatcher does the real routing
    Technicians accept the classification in the ordinary caseReclassification on pickup is routine and unmeasured

    Related criteria

    • AI ticket resolutionClosing five per cent of tickets with AI is an improvement. Closing most of the Tier 1 volume is a different business model. This criterion is about which one you are running.
    • AI-maintained documentationEvery MSP has documentation. The question this criterion asks is whether it describes the environment as it is today, and what keeps it that way.
    • Client transparencyDisclosure is now the criterion with a regulatory floor underneath it, and the floor is lower than what a client will actually ask for.

    All seven are listed on the criteria hub, and summarised on What is an AI-Native MSP?

    Browse the index