How MSPs onboard a new client's service desk in a day

Why does onboarding eat MSP margin?

The pattern is familiar. You sign a new client, and then the real work starts: workshops to map their services, spreadsheets of request types, weeks of back-and-forth about categories and priorities before anyone raises a single ticket properly.

During that gap, requests arrive anyway — by email, by phone, to whoever the client happens to know. Your engineers handle them without structure, nothing is categorised, and the SLA clock is a matter of opinion. The client’s first impression of your service desk is the messiest phase you will ever show them.

The hidden cost is the workshop time itself. Building a service catalog by committee burns the hours of exactly the people you hired to do technical work.

Generate the catalog from the service agreement

ReQuest DESK’s AI Builder inverts the process. Feed it the client’s service agreement, existing documentation, or even a plain-language description of what you do for them, and it generates a draft catalog: services, request types per service ranked by probability, and suggested handling for each.

Instead of starting workshops from a blank page, you start from a concrete draft and spend a single session correcting it. Reviewing is dramatically faster than authoring, and it is work an account manager can lead rather than an engineer.

Each step of each request type carries its own autonomy setting — AI, human, or an approval gate — so password resets can run hands-off while anything touching the client’s billing or infrastructure changes waits for sign-off. Nothing goes live until you publish it, and drafts stay drafts until you do.

One deployment, every client

Multi-tenancy is where this compounds. The multi-catalog registry runs any number of client catalogs from a single deployment, so onboarding client fifteen costs the same operational effort as client two.

In practice that means:

There is a fuller picture of the MSP setup on the MSP & IT teams page.

• Each client gets their own catalog, knowledge base and branded portal — it looks like their service desk, run by you.
• One classifier handles email, portal, chat, WhatsApp and transcribed voice across all clients.
• Every request is logged with an owner, timestamps and a committed resolution time, so SLA conversations are grounded in data.
• The proposals engine watches real cases per client and suggests new or changed request types — a human approves before anything changes.

Classification is your best sales demo

There is a commercial bonus. Because the AI Builder can draft a catalog from a prospect’s documents in a session, you can walk into a sales call with their service desk already sketched — paste in a realistic request and let them watch it classify with a confidence score and route to the right queue.

That is a far stronger close than a slide deck, and it is honest: what they see is what they get. With go-live measured in days — often a single working day — the gap between “yes” and a working desk stops being the riskiest part of the deal. If you want to see it against your own service agreement, request a demo and bring one.