Service 02
Most companies cannot name the AI tools their staff use.
AI use on corporate devices went from 15 percent of workers to 45 percent in a single reporting year, and two thirds of that use runs through personal accounts. An assessment tells you what is actually happening inside your business before you write a policy about it.
What an AI assessment covers
The assessment answers five questions: which AI tools are in use and by whom, what they cost across licenses and API spend, what company data is reaching them, which repeating work is worth automating, and where the governance gaps are. It runs two to three weeks and ends in a written findings document you can act on with us or without us.
- Inventory. The AI tools actually in use across the business, including the ones bought on a personal card and the browser extensions nobody approved.
- Spend. What you are paying in licenses, seats, and API usage, and how much of it maps to people who use the tool weekly.
- Data exposure. Which systems and documents those tools can reach, and which of them are covered by an agreement that protects you.
- Automation candidates. The repeating processes where automation would pay for itself, ranked by effort against return.
- Platform fit. Whether the tools you are paying for are the right ones for the work, which sometimes means using fewer of them.
The numbers behind this
Three findings from Verizon's 2026 Data Breach Investigations Report explain why this assessment exists. AI use on work devices tripled in a year, most of it is signed in personally, and unapproved browser extensions are now common enough to assume you have them.
Which platforms we work with
We separate the tools we implement from the tools we find. Implementation means we will design, build, and support real work on that platform. Discovery means we will find it in use, assess what it touches, and bring it under policy. Claiming implementation depth on every AI product on the market would not be true, and you would discover that during delivery instead of during a sales call.
| Platform | What we do | What we typically find |
|---|---|---|
| Microsoft 365 Copilot | Implement and support | Already licensed and underused. See Copilot enablement |
| ChatGPT and the OpenAI API | Implement and support | Team or Enterprise plans alongside personal accounts nobody tracks |
| Claude and the Anthropic API | Implement and support | Document analysis and drafting, and the strongest fit for custom build work |
| Google Gemini and Workspace | Implement and support | Workspace businesses using Gemini inside Docs, Sheets, and Gmail |
| Perplexity, Notion AI, Grammarly, Cursor, and the rest | Discover and govern | Adopted by individual staff, almost never on a company agreement |
What it costs
An AI assessment starts at $2,000 as a fixed fee, and the figure is the same whether the findings are comfortable or not. Larger or messier environments cost more, and we tell you which one you are before starting. The assessment is deliberately priced so it can be bought without a committee, because the point is to replace a guess with a number.
What happens after the assessment
The findings point in one of three directions, and often more than one. If the answer is that you already own the right tool and nobody uses it, that is Copilot enablement. If the answer is that a specific process is worth automating, that is workflow automation. If the answer is that data is reaching places it should not, that is AI governance. Sometimes the answer is to do none of it yet, and we will say so.
What we will not tell you
We will not tell you that you need more AI. An assessment that always concludes with a larger purchase is a sales process wearing a report as a disguise. For a business paying for seats nobody opens, the honest recommendation is usually to cut the seat count and enable the people who remain.
Questions
Common questions
What is an AI assessment?
A fixed-fee review of the AI tools your business already uses, what they cost, what data they can reach, and which repeating processes are worth automating. It takes two to three weeks and produces a written findings document. It is vendor-neutral and covers Microsoft Copilot, ChatGPT, Claude, Gemini, and the consumer tools staff adopt on their own.
Do you only work with Microsoft Copilot?
No. Copilot gets specific attention because most businesses are already paying for it, but we implement and support ChatGPT and the OpenAI API, Claude and the Anthropic API, and Google Gemini. Tools like Perplexity, Notion AI, Grammarly, and Cursor are covered on the discovery and governance side, where they are usually found in use without an agreement.
How do you find AI tools nobody told you about?
Through DNS and network telemetry, browser extension inventory, tenant and identity sign-in logs, and expense review. Most discoveries come from the last one, because unapproved tools are usually paid for, just not by IT.
What does an AI assessment cost?
Assessments start at $2,000 as a fixed fee, covering inventory, spend analysis, data exposure review, and a ranked list of automation candidates. Implementation and remediation are quoted separately once the scope is known.
We are a Google Workspace company, not Microsoft. Is that a problem?
No. The assessment is platform-neutral and covers Gemini and Workspace the same way it covers Microsoft 365. Our managed IT practice is deepest on Microsoft, and we will tell you plainly where that changes what we recommend.
Can you help if we have not adopted any AI yet?
Yes, and that is a cleaner starting point. The assessment then focuses on what your staff are already doing informally, which processes would benefit first, and what has to be true before a rollout is safe.
Next step
Thirty minutes, and you will know where you stand.
No pitch deck. We look at what you are running, tell you what we would do first, and you decide whether that is worth paying for.