A free 15-minute Zoom call. No cost, no pressure.Why IT firms put off their own AI setup
You know what AI is. You have probably tested it on a personal account, gotten a decent answer, then closed the tab and gone back to tickets. The problem is not interest. It is that setting it up right for a whole team takes an afternoon you do not have, and doing it half way means your techs use it in five different ways with five different results.
There is a second reason. You are careful with client data because that is your job. So you want to know what goes where before you roll anything out. That is fair, and it is exactly the conversation to have before setup, not after.
We are not here to replace your stack or sell you software. We set up Claude, ChatGPT or Gemini on your own account, connect it to the tools you already use, and train your people. You keep paying the AI company directly.
Start with the writing work, not the tickets
The fastest win for an MSP is the writing that happens after the fix. Ticket replies, resolution notes, documentation, client reports, proposals and scope-of-work drafts. It is repetitive, it is text, and it is where AI is strongest.
A good first week looks like this:
- Set up the account and get every tech logged in on their own profile.
- Connect it to your email and shared files so it can see the ticket thread and the client's setup notes without anyone pasting them in.
- Load your tone and your templates. Short, plain, no fluff. Your clients should not be able to tell which reply was drafted.
- Have each tech run it on real tickets for a few days and edit what comes back.
Once replies and notes feel normal, move to documentation and reports. Proposals come after that, because they need more of your judgment in the prompt.
Documentation is where the hours hide
Every MSP has a folder of runbooks that are half written, out of date, or living in one senior tech's head. AI is good at turning a rough set of steps into a clean document, and good at rewriting an old runbook when a client's environment changes.
A practical pattern: the tech does the work, then talks or types the steps out rough, and AI shapes it into your standard format. Same for onboarding notes on a new client, network summaries, and the plain-English explanation you send a client after a change.
Watch out for one thing. AI will fill a gap with a confident guess. If a runbook step is wrong, it will not always flag it. Have the tech who did the work read the final doc before it goes into the shared folder. Treat the draft as a draft.
Client reports and proposals
Monthly reports and proposals eat evenings. Both are mostly structure plus your specifics, which is exactly what AI handles well once it has your material.
For reports, feed it the ticket and project activity for the month and your usual report layout. It drafts the summary, the highlights and the watch items. You check the numbers and the tone, then send. The report goes out on time instead of the following week.
For proposals, keep a folder of past ones. AI can match the shape of a good proposal to a new scope, and it is useful for the part everyone skips, which is spelling out what is not included. That one habit helps keep scope creep in check.
Be clear with your team that AI drafts, the account manager decides. A proposal with a wrong assumption in it still costs you the deal.
Client files, zero data retention and private AI
Your clients trust you with their systems, and some of them will ask what you are doing with AI. Have an answer ready.
If you handle sensitive client files, there are two options. Zero data retention means the AI provider does not keep your prompts or files after it answers. That is not the same as not used for training, so say it correctly. The other option is private AI on a machine in your office, so files never leave the building.
For your own shop, a sensible rule is: general writing can use the standard account, anything with client identifiers or config detail goes through the zero data retention setup or the private machine. Write that rule down so every tech follows the same line. It also becomes a selling point when a prospect asks how you use AI.
What a normal week looks like after setup
Monday, the techs draft ticket replies and resolution notes with AI and edit them before sending. Tuesday, one person updates the runbooks that drifted. Wednesday, the account manager drafts the monthly client reports from the activity log. Thursday, a proposal goes out that started as a rough scope. Friday, someone asks a question in the support thread and gets an answer the same day.
It is not dramatic. It is the same work, finished with less drag, and your evenings come back.
Two things to watch. First, do not let AI answer a technical question without a tech checking it, especially on a live client issue. Second, watch for personal accounts creeping in. If half the team is on a free login, you lose the shared files and the consistency. Get everyone on the company account in the first place.
Every client starts with Setup, then can move up to Support, Coaching or Done for you whenever it makes sense. If you want automations, an AI phone assistant or private AI in your office later, that is there too.
How to start
Book the free AI game plan. It is a free 15-minute Zoom call, no cost and no pressure. We look at your firm, point at where AI helps first, and tell you which level fits. Nothing gets set up on that call.
After that you get your price in writing before anything starts, and we begin with Setup. You keep your own AI account the whole way, and if you ever stop, the account stays yours.
Book the free AI game plan and we will map the first three things to set up for your team.