Your clients are going to buy AI from someone. Make it you.
Breeze is an RMM: one agent on your clients’ machines, and a ready-made AI practice on top of it. Assistants their employees use every day, delivered under your brand, governed by policy you control. You already have the trust, the endpoints, and the billing relationship. Now you have the platform.
Your brandWord / Excel / PowerPoint / OutlookGoverned by policyRecurring revenue
Start Cloud Beta opens the signup form on our home page. If you would rather ask questions first, book the call: we walk the policy model with you and you keep the answers either way.
The moment in front of you
Why handing out licenses is not an AI practice.
The AI Conversation Is Happening Without You
Your clients are already asking about AI. If the answer comes from a generic AI consultant or a software reseller, you lose the most strategic conversation you have had with them in years, and someone else owns the roadmap.
Copilot Licenses Write You Out of the Story
Reselling licenses hands your client a chatbot and hands you nothing: no ongoing role, nothing of your own to configure or audit, no recurring revenue. When the license renews without you, you were never part of the value.
Shadow AI Is Already Inside Their Business
Their employees are pasting company data into public chatbots today. Someone has to put a governed alternative in front of those users before the first incident, and that someone should be you.
You do not need to build an AI practice from scratch.It ships inside the RMM. If you already run Breeze, it is a policy you enable for one client. If you don’t, it comes with the platform the day you deploy it.
What your client sees. What you control.
One request inside a client’s spreadsheet, shown from both ends at once. The policy keys on the left are the same keys you set on the right.
I’ll read the sheet first, then propose the column as a change you approve.
read_rangeran on its own
E14redacted · financial
That cell held a card number. DLP replaced it before the model saw the sheet, and the log keeps the redacted form. dlpConfig
write_range · Sheet1!G2:G214waiting for approval
Cell
Before
After
G2
empty
REVIEW
G3
empty
ok
G4
empty
REVIEW
3 of 213 rows shown.
ApproveReject
Nothing touches the workbook until Dana clicks approve. Auto-apply is off for this org. writeApproval
Ask about this workbook
The same session, from your side
Breeze console · AI for OfficeYou
Org policy · Riverbend Dental
Policy keys (8)
branding
Northline IT, your logo
writeMode
set to readwritereadonly
writeApproval
set to askallow_auto
userAccess
allset to selected
dlpConfig
financial: redact · credential: redact
dailyBudgetCents
2500
perUserMessagesPerMinute
10
retentionDays
30
You set this per client organization. Breeze scans their endpoints for the same classes of data it redacts here: see sensitive data discovery.
How this turn was classified
read_rangenon-mutatingruns immediately
write_rangemutatingparked as a preview
Enforced on the server, not in the pane. Switch writeMode to readonly and the write tools are stripped from the model’s toolset before the session starts. The same tiering governs what your technicians’ AI can do: AI Risk Engine.
Every prompt, tool call, redaction, and approval is recorded. When your client asks what the AI did in their books, you open this and show them.
Illustrative depiction of Breeze AI for Office. Northline IT and Riverbend Dental are example names, not customers, and the log entries are made up. The branding slot, the DLP redaction step, the before/after approval card, the policy keys, and the audit fields are how the product actually works.
Want to see this on your own tenant before you commit to anything? Book a call and we will walk the policy model with you.
Four surfaces, one set of controls
The pane above is one of them. Each one answers to the same org policy, the same DLP step, and the same audit trail.
WorkspaceIn developmentA dedicated environment where AI will work alongside your clients' teams on real files.
A practice, not a project
Start with one client. The platform does not make you commit to a transformation.
1
Start with one client
Enable AI for Office for a client who is already asking about AI. Set the policy conservative: read-only mode, or previews on every change.
2
Put your brand on it
The pane carries your name and logo through white-label branding. Their employees see their IT partner delivering AI, not a third-party logo.
3
Charge for it
Quotes and billing live in the same platform. Package AI as a monthly service line, not a one-off project invoice.
4
Expand the practice
Roll Breeze Helper onto desktops, enable computer control for operations work, and manage every client from one console.
Bottom lineRecurring revenue, not a one-off engagement.And the governance stays yours the whole way.
What it costs you. What you charge.
A service line needs numbers, so here are the only ones we can honestly source: ours.
The platform
AI for Office is part of Breeze, not an add-on. All 64 modules ship in every plan, with no per-device fees, no per-technician seats, and nothing paywalled. Turning the practice on for a client is a policy you enable on their organization, not a SKU you buy.
What Breeze bills you
Self-hosted Community is free under AGPL-3.0, with no device limit and no seat count. Breeze Cloud is flat tiers in public beta: $20/year up to 3 devices, $99/month up to 250. Never metered per device or per seat. Full pricing.
What the AI costs
Model usage is the one variable, and it runs on your own key. Bring your own Anthropic API key on any plan and you pay the provider directly, at their price. The $99 Cloud plan includes credit-based Breeze AI instead, if you would rather not manage a key at all.
Your ceiling
Set per client organization, in the same policy that carries your logo: a daily budget, a monthly budget, a per-user message rate limit, and an hourly cap for the whole org. A client cannot run up spend you did not authorize, because the cap is enforced on the server, not in the pane.
What you charge
Yours to decide. We do not set your price and we do not sell past you. Quotes and billing are modules in the same platform, so the AI line lands on the invoice you already send that client every month.
We are not going to print a suggested retail price here. We do not yet have enough MSPs billing this practice to publish a number we could defend, and an invented one is worse than none. Bring the client and the seat count you have in mind and we will walk the cost floor with you: book a call.
Start with one client, one policy, and read-only mode. Deploy Breeze, turn the pane on for the client who is already asking, and widen the practice when it has proven itself.
Start Cloud Beta opens the signup form on our home page. The call is the faster route if you want the policy model and the economics walked through first.
Still evaluating Breeze?
Get your questions answered by the team that builds it. 20 minutes, no deck.