Common questions.
Do you ever see our data?
No. Your documents, the questions your team asks, and the answers they get back stay on your equipment, inside your building. Our maintenance systems collect health and performance metrics only, never your content. That’s written into the agreement, not just configured.
Where does the AI actually run?
On hardware you own, on your own network. There’s no third-party cloud in the loop. That’s the whole point. It’s why it’s called Clear Box.
Isn’t building our own AI expensive?
There’s an upfront cost for the hardware, which you own. But you’re buying an asset, not renting a meter. At real usage, owning the equipment can be dramatically cheaper than paying a cloud provider per token, often paying for itself in months. We put the numbers in front of you during the assessment, so you decide with a real comparison, not a guess.
How is this different from ChatGPT, Copilot, or cloud enterprise AI?
Those send your data to someone else’s servers. For a firm whose confidentiality obligations are the business, that’s the problem, not the product. We build the same kind of capability on infrastructure you control, so your data never leaves.
We don’t have anyone technical to run this. Is that a problem?
No. That’s why Managed Care exists. We build it, and we keep it running: monitoring, patching, updates, and support. You don’t need ML engineers on staff.
What if we want to leave, or bring it in-house later?
It’s your hardware and your system. Offboarding is documented up front and takes about a day: we hand over the configuration, credentials, and a final backup, and revoke our access. No lock-in.
How long does it take to get running?
Weeks, not quarters. The assessment scopes the build precisely, so there are no open-ended timelines.
What does the whole thing cost?
A fixed fee for the assessment, a quoted price for the build (hardware billed at cost), and a monthly Managed Care plan. Everything is quoted up front. No open-ended or metered billing.
Is this a fit for our kind of organization?
If your confidentiality obligations rule out sending data to the cloud, yes. That includes privileged legal files, client financial records, protected health information, and any business whose intellectual property is the business. The assessment confirms the fit before you commit to anything.
Can our IT team or security reviewer vet you first?
Please do. We expect it, and we’ve designed everything to make the review straightforward. We’re happy to walk your team through the architecture in detail.
What kind of AI can it actually do?
The common starting point is a private assistant that answers questions grounded in your documents (contracts, matter files, records) instead of the open internet. We scope the specific use cases to your firm in the assessment.