Start a client in ten minutes.
Your client's administrator authorizes the read: about ten minutes. You leave with an assessment in your name.
What you ask yourself
- Will my client agree to give access?
- What has to be installed?
- What do I hand over after the readout?
- And if they are on Google Workspace?
What Ojja gives you
- A read-only authorization link. Your client's administrator authorizes Ojja's application, with an administrator account. The exact list of permissions is published. Nothing to install on workstations.
- Three counts. AI connected by their employees to their data; AI turned on by their vendors with no decision; AI that trains its models on their data or sends it outside the EU.
- An assessment in your name. You hand it to your client, with your reading.
- A twenty-minute presentation. The outline from the kit: how to present the three counts, and what comes next.
- Then the register, with Ojja Watch. One line per feature, agent or application, with its origin, its usage read over 28 days, its cost and its data.
The readout reads Microsoft 365 today. The rest is declared by your client, and each line says where it comes from.
On the Arvenne example
Arvenne's readout reads 412 accounts, 37 applications and 14 AI features. Its three counts: 7 AI features connected by employees to its data, including ChatGPT on 14 accounts; one AI feature turned on by a vendor with no decision, the meeting recaps; 2 AI features that train their models on its data or send it outside the EU, the image generator and the meeting transcription tool. That is the assessment you present.
Arvenne, a fictional company · sample data
What Ojja never does
- Ojja reads registers and configurations, never your clients' content.
- Ojja never writes in your client's tools: it does not turn on, turn off or set anything.
- Ojja gives no figure per person.
- After the readout is handed over, Ojja reads nothing more, and your client withdraws access whenever they want.