Use AI on the work that actually matters.

Most small firms are quietly holding their best work back from AI, because nobody has told them where the data goes. We fix that without changing the tools you already use.

Book a 20-min intro call →or get a written quote →

·· 01 ·· A day in the life

The quarter you stop holding back.

Managed AI is not a tool we install. It is a set of decisions somebody has to own — which plan, which settings, which data — and then keep owning as the tools change underneath you. Here is what that ownership actually looks like on the calendar.

Week one

Finding out what is already in use.

Almost nobody knows. Someone in operations is on a personal ChatGPT plan, somebody in finance signed up for a trial, and Copilot came switched on with a licence nobody chose. We start by making that list, because you cannot govern a tool you did not know you had.

Deliverable · what is in use, in writing

Week two

The checkbox nobody checked.

On a consumer plan, conversations can be retained for years and used to train the model, and the setting that stops it is one toggle deep in an account nobody audits. We verify it, evidence it, and move you to the right plan — which, more often than owners expect, costs less than the one they are on.

Deliverable · settings verified and evidenced

The turn

A list of permissions, not restrictions.

The written data-class map is the piece that changes behaviour. It says what is safe to put in — and once that exists in writing, the work people were quietly avoiding starts moving. The client file. The messy financials. The proposal that would have taken all afternoon.

Deliverable · a written data-class map

Quarterly

Because the tools keep changing.

Vendors ship new features, new connectors, and occasionally new defaults. So the training setting gets re-verified, the attestation gets refreshed and signed, and a new AI service appearing on your network raises an alert rather than a surprise a year later.

Deliverable · signed attestation, refreshed quarterly

Your team is already using AI. The only open question is what it is doing with your clients’ information.

Why this is governance, not a product rollout

What we watch — and what we don’t

We govern where your information travels. We do not read what anyone types.

Where the data goes, evidenced. What is in the prompt, never.

PlanTerms, in plain language — what your subscription actually permits the vendor to do, what it costs, and what the correct plan costs instead.

ConfigTraining settings and single sign-on — verified, evidenced, and re-checked, rather than assumed because somebody remembers switching it off.

ConnectorsScoped deliberately — the AI reaches the Microsoft 365 content that helps and nothing beyond it.

Not usPrompt content. Content inspection is not available below an Enterprise plan, and we would not want it. We watch travel, not conversation.

·· 02 ·· The specification

The problem.

Your team is already using AI. The question is what it is doing with your clients’ information.

On a personal Claude or ChatGPT subscription, conversations can be retained for five years and used to train the model. Turning that off is a checkbox — one nobody verifies, and one that sits outside every system your IT provider manages.

Meanwhile the work that would benefit most from AI is the work nobody dares paste in.

What we do.

Audit the plan you are on

What your subscription’s terms actually say, what it costs, and what a better plan would cost. Usually less than you expect.

Configure it properly

Training settings verified and evidenced. Single sign-on. Connectors scoped so AI reaches what helps and nothing else.

Tell you what is safe to put in

A written data-class map. Not a list of restrictions — a list of permissions, so the work you have been avoiding can start.

Monitor it and prove it

Ongoing visibility into what AI is in use across your business, and a signed statement of where your data goes that you can hand to a client who asks.

What it works with.

Claude, ChatGPT, Microsoft 365 Copilot, and Gemini. Keep the subscription you have. If a better plan costs less — and it often does — we will show you the numbers and handle the move.

What’s not included.

We do not build AI applications, agents, or integrations, and we do not replace or reimplement the tool your team already uses. Model fine-tuning, training on your data, and locally hosted models are outside this service. We monitor where data goes, not what anyone types: prompt-level content inspection is not available below an Enterprise plan, and we do not read prompts. We govern where your information travels — we do not warrant the accuracy of what a model says.

Service-level commitments

Initial response (managed clients)≤ 1 business hour

Training setting re-verifiedQuarterly

Attestation refreshedQuarterly

New-service alertingIncluded

Quarterly posture reviewIncluded

·· 03 ·· Plans

Four tiers, priced per environment.

Essential

$299

per month · up to 10 users · 2 hrs / qtr

Small team, mixed tools, keeping their subscriptions. Plan and terms assessment, training settings verified and evidenced, connector scoping, monitoring, and a written acceptable-use policy.

Standard

$599

per month · up to 25 users · 4 hrs / qtr

Unified business tenant, one primary AI vendor. Adds tenant configuration and single sign-on, connector hygiene, the written data-class map, and a signed attestation refreshed quarterly.

Advanced

$999

per month · up to 50 users · 8 hrs / qtr

Multiple vendors, staff-wide use, role-based data classes. Adds Claude, ChatGPT, and Copilot governed in one map, per-role permissions, staff enablement, and an AI clause for your own client contracts.

Complete

$1,699

per month · up to 50 users · 12 hrs / qtr

Maximum available control, and oversight of an Enterprise-plan tenant where you have one. Adds monthly audit-log review, SCIM lifecycle, role-based access design, spend governance, and retention set to your own contractual commitments.

Choose the tier by the protection you want, not by your headcount. A two-person firm handling sensitive client material can take Complete. A twenty-person firm with simple needs can take Essential.

Users above the included count are $29 each per month on Essential and Standard, $25 on Advanced and Complete. The bands are set so the next tier costs less before the overage does — Standard is cheaper than Essential from 21 users, Advanced cheaper than Standard from 39. Term and annual-prepay discounts apply as they do on every plan. Where your current subscription cannot support something a tier includes, we say so in writing and price the path — we never drop it quietly.

·· 04 ·· Start here

Start with a free AI exposure review.

A written answer to the question your own clients will eventually ask you: where does my information go, and who can reach it. We look at what AI is actually in use, what your current plan’s terms say, what data is exposed, what the correct plan costs, and what closing the gap costs. No charge, and the first fix is usually free and takes minutes.

Your name ·required

Organization ·required

Email ·required

Which AI tools are in use?

What plan are you on?

How many people use it?

Phone

What are you holding back from AI, and why?

Your info stays with us · we don’t sell or share leads

Microsoft Cloud Partner Google Workspace Partner Kaseya / Datto MSP Apple Business Partner Dell Authorized Partner Lenovo Partner Ubiquiti Certified Installer HIPAA-Aligned SOC-Aligned CJIS-Aligned PCI-Aligned