Call us on 1300 360 369 Customer Portal

AI Managed Services

Horizon AI365

A managed service that governs Microsoft 365 Copilot, measures whether it is working, and keeps it improving month after month. Scoped to Copilot, delivered remotely, and built for Australian organisations running roughly 20 to 500 seats. It is advisory and governance led: we assess, prioritise and recommend, and your team stays in control of the tenant.

You bought the licences. Nothing changed.

  • 82% of leaders say this is the year to rethink strategy and operations around AI
  • 80% of the global workforce report more work to do than time or energy to do it
  • 2 min the average gap between interruptions in a knowledge worker’s day

Microsoft Work Trend Index Annual Report 2025, from 31,000 workers across 31 countries.

It grows whether or not anyone is governing it

A few people try Copilot. More follow. Agents start appearing and nobody is certain how many there are or who owns them. At some point it has a shape. The only question is whether that shape was decided or whether it just happened.

  • Scattered individual use

    Licences are active, usage is sporadic, governance is informal and the value is anecdotal. No agents yet. The most common place to start, and the hardest to measure later if nobody sets a baseline now.

  • Clusters, and the first agents

    Six to eighteen months in. Pockets of real habit have formed and the rest of the organisation has none. The first agents appear, usually without anyone having approved them.

  • An environment with a boundary around it

    Copilot sits in priority workflows, agents are throughout, and the whole thing is inside something you can account for. Fewer organisations are here. It is the only state where anyone can answer what AI is doing in the business.

What this is, and what it is not

Horizon AI365 is advisory and governance led. We look, measure, prioritise and recommend. Your team executes. Being blunt about that line is what makes the rest of this page worth reading, and it is what stops the monthly fee quietly turning into something nobody scoped.

What it does

  • Governance you could show a board

    An AI Council with real decision rights, an AI usage policy, a Responsible AI charter, a value case, and a monthly forum where priorities are set and decisions are written down.

  • Adoption measured, not assumed

    Usage reviewed against the scenarios you said mattered. Persona and journey mapping, champion network design, wave onboarding patterns, and adoption measures that turn anecdotes into evidence.

  • Readiness assessed against Microsoft’s own framework

    Tenant baseline across identity, data, security and compliance. Sensitivity labels, data loss prevention and Purview readiness guidance, and a prioritised list of what to fix first.

  • A rhythm that does not depend on you chasing it

    Plan, Do, Review and Evolve every cycle, with minutes, decision records, an action register and a risk register kept current. A quarterly executive rollup every third cycle.

What it does not do

  • Not a service desk

    No incident response, no user issue resolution, no Copilot support tickets. There is no technical escalation path into this service. That stays with your IT support function.

  • Not remediation

    We identify and prioritise what needs fixing. We do not execute the fix. Recommendations you choose not to act on remain your risk, and we will keep saying so on the record.

  • Not tenant changes

    No configuration and no change control. Access is read level, provisioned with your IT and security teams at onboarding and revoked when the engagement ends.

  • Not training delivery

    We design the adoption approach, the champion network and the learning patterns. Running the sessions and building the content is project work, scoped separately.

  • Not agent building

    Complex agent design, build and deployment, custom connectors, integrations and Power Platform development are all projects. We govern agents. We do not build them inside the managed fee.

  • Not every AI tool you own

    Scoped to Microsoft 365 Copilot. Not ChatGPT Enterprise, not Gemini, not Claude. Broader scope is a separate piece of work rather than something we quietly absorb.

What each tier buys

All three workstreams run in every tier. The tier decides how deep each one goes, and it reflects engagement intensity rather than your size or your maturity: you can sit at any tier at any of the three states above.

What each tier includes, by workstream. All three workstreams run in every tier; the tier decides how deep each one goes.
Foundation $2,330 8 advisory hours Accelerate $4,660 16 advisory hours Scale $9,320 32 advisory hours
Leadership and governance The mandatory base of every engagement, and the one that sets the terms for the other two. It owns decision rights, escalation and prioritisation: what is allowed, what is funded, and what gets measured. Outputs are a live AI Council, a usage policy, a monthly review pack and governance measures that survive an audit.
Governance forum Monthly Monthly sponsor and service review Formal monthly cadence across leadership, IT, security and change
Executive alignment and briefing Light Moderate Deep
Decision records, action and risk registers Included Included Included
Scenario prioritisation and roadmap decisions Light Active Full
Agent and portfolio governance Not included Not included Included
Human change and adoption Shifting behaviour rather than enabling software. Persona and journey mapping, communications planning, champion networks, wave onboarding, manager coaching, and role descriptions rewritten as the work changes. Advisory: it does not include running training or change campaigns.
Usage review against your priority scenarios High level Active Deep
Adoption gap identification Light Active Comprehensive
Advisory adoption recommendations Included Included Included
Delivery module execution Not included One per cycle Several per cycle
Champion network and enablement advisory Not included Via delivery module Ongoing
Technical readiness and operations Visibility, assurance and risk awareness across the tenant. Baseline and posture audit, sensitivity labels, data loss prevention, Purview readiness, workload integration patterns and lifecycle advisory. Strictly advisory and read only. It sets the floor for what is safe and bounds how far adoption should go.
Copilot telemetry and service health review Included Included Included
Technical readiness assessment Light Active Deep
Readiness gap analysis Not included Via delivery module Ongoing
Security and compliance oversight Light Moderate Full
Measurement reporting and benchmarking Baseline Detailed Comprehensive

Australian dollars, per month, on a six month initial term and then three monthly. A one off onboarding fee of $2,330 covers the kick off and the preparation of your programme assets, roughly a day of effort across a week. Additional hours inside the service boundary are $340 each by written agreement. Hours are capacity rather than a bank you draw on: work is prioritised through the governance forum, and unused time does not roll over unless we agree it in writing. Tier changes take effect from the next billing cycle. Annually that runs from about $28,000 at Foundation to about $112,000 at Scale.

What you keep if you walk away

  • The governance environment

    A SharePoint site holding every register, report, decision record and policy, and a Teams team for the work between sessions. Neutrally branded by default, with no logos of ours on it. Yours from day one, not at the end.

  • The documents

    AI usage policy, Responsible AI charter, AI Council charter, value case, risk and issue register, readiness scorecard, maturity history and your agreed measures. All of it stays.

  • Our access, removed

    We hold external guest access configured against your conditional access posture, and it is revoked at off-boarding by a documented process. The environment keeps working without us in it. That is the point of building it in your tenant.

How a month runs

A rolling 30 day cycle from your start date, not the calendar month. Four phases, the same in every tier. What changes between tiers is depth and hours, not the shape.

  1. Plan

    Week one, mostly ours. We confirm the priorities for the cycle, revisit the scenario register, and decide which of the previous cycle’s recommendations you are endorsing, deferring or rejecting.

  2. Do

    Weeks two and three, where you see us most. The governance forum runs, decisions get made and recorded, and advisory work lands across all three workstreams. On Accelerate and Scale this is where the cycle’s delivery module is delivered.

  3. Review

    Week three. Usage and adoption telemetry, the risk register, adoption quality against your priority scenarios, and where you now sit on the maturity curve. Reporting is informational and advisory. No service levels apply to it.

  4. Evolve

    Week four, ours again. We reassess maturity, log the optimisation opportunities, close the cycle on the record and set up the next one. Nothing carries over silently.

Managed intelligence, not managed infrastructure

Keeping systems running and keeping AI capability useful are different jobs, and they need a different kind of month. Copilot is not a deployment that finishes. It is a capability that either matures or quietly stops paying for itself, and almost nobody in the mid market is being paid to care which.

What we need from you

This works when four things are true. They are worth reading before the price, because the fee buys the same thing whether or not they are in place, and without them nothing moves.

  • An executive sponsor

    Somebody accountable for AI outcomes who turns up to the governance forum. Without a sponsor the recommendations stack up and none of them get decided.

  • A service owner

    One person on your side who coordinates inputs, decisions and access. Not a full time job. It is the role that makes the month run.

  • Read access to your Copilot telemetry

    Copilot Dashboard, audit log read, the right admin roles, and Purview where a posture review is in scope. Provisioned at onboarding after a security review with your IT and security teams.

  • Decisions that get actioned

    We prioritise and recommend, you execute. Recommendations that are not acted on stay on your risk register rather than ours. That is the honest shape of an advisory service and it is better said now than in month four.

Common questions

Short answers first. If something here is a deal breaker, better to find out now than in the third meeting.

What is a Copilot readiness assessment?

A structured review of whether your Microsoft 365 tenant, your data and your organisation are ready for Copilot to be useful and safe. It covers four things: technical posture across identity, data, security and compliance; data readiness, meaning sensitivity labels and data loss prevention, so Copilot does not surface documents to people who should not see them; governance maturity, meaning whether anyone owns AI decisions; and an adoption baseline you can measure later improvement against.

Ours is a facilitated workshop against a Microsoft aligned framework. It produces a findings report, a scored position, and a prioritised list of what to fix first. It is the first thing that happens in an engagement, and it is the reason the second month is not guesswork.

How much does Microsoft 365 Copilot cost in Australia?

The Copilot licence is bought from Microsoft or a Microsoft partner and priced per user per month. Microsoft publishes the current Australian figure and changes it from time to time, so check it there rather than trusting a number on a page like this one.

Our fee is separate and is not a licence resale. Horizon AI365 starts at $2,330 a month for the managed service and assumes you already licence Copilot for the users in scope. If nobody is licensed yet there is nothing to govern, and we will say so.

We already have Copilot licences. What does this actually do?

It closes the gap between having Copilot and getting something out of it. Most organisations with licences activated them and then saw sporadic, ad hoc value, because governance stayed informal, adoption stayed uneven and nobody built a business case anyone could check. This is the recurring service that sustains it: a governance cadence, adoption measured against the scenarios you care about, and readiness assurance, without you having to hire the capability.

Is this just consulting with a subscription wrapper?

No, and the difference is testable. Consulting is project shaped: scope, plan, deliver, leave. This is a recurring monthly rhythm with a fixed capacity of hours and a boundary we will point at in writing when something falls outside it. You are not buying a deliverable, you are buying sustained capability. The two work together, and when execution is genuinely needed we scope it as a project rather than pretending the retainer covers it.

Does it include training?

No. Adoption advisory is included: champion network design, wave onboarding patterns, role based learning patterns and adoption measurement. Delivering training sessions and building training content is project work. Drawing that line is what keeps the monthly fee honest.

What if we need more than the included hours?

Additional hours inside the service boundary are $340 each by written agreement. Work that sits outside the boundary, such as configuration, remediation, training delivery or complex agent build, is scoped separately as a project. Mid cycle scope increases are held to the next governance checkpoint rather than absorbed silently.

Can we move between tiers?

Yes, by agreement, effective from the next billing cycle rather than mid cycle. It does not need a new contract. Most organisations start at Foundation to get governance and visibility in place, then move up when there is something specific worth going deeper on.

What is shadow AI, and does this cover it?

Shadow AI is staff using AI tools the organisation has not approved and cannot see: pasting client information into a free chatbot, running documents through a browser extension, using a personal account for work. The risk is already present in most organisations, whether or not anyone has looked.

Bringing it under control is the governance workstream's first job: an AI usage policy that says what is and is not allowed, data handling standards, and a forum that keeps both current as the tools change. Detecting and blocking specific tools is a security control rather than an advisory one, and it belongs with your security stack.

What if Microsoft changes Copilot significantly?

The service is built on Microsoft’s Copilot Adoption Playbook and Success Kit rather than on a framework of our own, so when Microsoft moves, we move with it and you are not exposed to us rebuilding our method. Reporting reflects whatever telemetry Microsoft exposes, which does change, and we will tell you when a measure stops being available rather than quietly dropping it from the pack.

How do you know whether it is working?

Three levels, agreed at onboarding and reviewed every cycle. Governance maturity: are decisions being made and recorded, is policy iterating, are risks being managed. Adoption quality: usage patterns against your priority scenarios, champion network activity, Copilot Dashboard signals. Value realisation: outcomes against the value case you endorsed. All of it is informational and advisory, and none of it carries a service level.

Start with a Copilot readiness assessment

Onboarding is about a day of effort across a week. It runs before anything else because it produces the baseline everything afterwards is measured against, and that baseline gets harder to establish every month you wait.

  1. Readiness and maturity assessment

    A facilitated workshop against a Microsoft aligned framework, covering tenant posture, data readiness, governance maturity and your adoption baseline. You get a findings report, a scored position, and a prioritised list of what to address first.

  2. First planning session

    We agree the priorities for cycle one, the scenarios that matter, who sits on the AI Council, and the measures we will report against. That becomes your success plan and the reference point for every cycle after it.

  3. Your governance environment, provisioned

    SharePoint site, Teams team, registers and starter policies in place, and advisor access configured with your IT and security teams after a pre onboarding security review.

Find out where you stand

A readiness assessment is the honest first step. It tells you what is in the way, what to do about it first, and whether a managed service is the right shape for you at all. Sometimes the answer is that it is not.

Book a readiness assessment