Questions owners ask us.

No — not the way you're picturing. We rent the engine from the big AI companies on their business terms, and those terms say your information is never used to train their AI, and the conversation isn't kept by them. Here's how it actually works: Odell is like a sharp employee with keys to your filing cabinets. Somebody asks a question, Odell opens only the cabinets that person is allowed to open, reads what's needed, answers — and that's the end of it. Your cabinets stay locked down and encrypted, and here's the part owners like best: every time anybody looks in a cabinet — an employee or Odell — it's written in a log that you can read. You hold the keys. Compare that to right now, where your prices live in a binder anybody can open and nobody logs a thing.

ChatGPT knows the internet. Odell knows your company — your prices, your procedures, your people's judgment — because it was interviewed out of them, organized into cabinets, and approved by you. It also respects your org chart's secrets: it answers each employee only from knowledge they're allowed to see, cites its source every time, and says “I don't know that yet” instead of improvising a confident wrong answer.

It says so — plainly, no bluffing. The question becomes a gap ticket routed to the person whose cabinet should know. When they answer and it's approved, that knowledge is in Odell for everyone who's allowed to see it. The brain grows exactly where it was thin, as a byproduct of being used.

The entry gets flagged. Odell leads any answer with the flag and treats the fact as unsettled, and only an approved person — the owner or the operator — can settle which document is right. The decision is written onto the entry itself and into the audit log. This isn't hypothetical: during our first pilot load, two manufacturer documents stated two different warranty terms. Odell flagged it instead of picking one, and it stays flagged until the owner settles it with the source.

You do. Interviews, uploads, and forwarded emails all produce proposed entries. Nothing is answerable until a human — the owner, or the owner of that cabinet — approves it. Documents are sources; approved entries are what Odell answers from.

Permissions are applied before Odell reads anything, not after. An employee's Odell can only open cabinets they're cleared for; knowledge outside them isn't summarized, hinted at, or reachable through clever phrasing. Before the first pilot touched real company knowledge, the permission engine passed an automated evaluation built from that company's own cabinets — every permission block, document gate, and cross-company check held, with zero leaks.

No. Models are called under paid, no-training business terms, with zero-retention settings wherever the provider offers them — and model families whose terms force retention near customer data are excluded from the roster entirely. Conversations are ephemeral: when a session ends, Odell forgets it.

That's exactly when to do it. The interview is a conversation, not homework — they talk, Odell organizes. And owners tell us the go-to person is usually relieved: the phone stops being their second job, and what they know stops depending on them showing up every day.

Yes. Ask for the paper — the bond rider, the insurance certificate, the spec sheet, the signed warranty — and Odell hands it over to view or download, under the same permissions as the knowledge in it. Every open and download is written to the log.

Voice is a bolt-on module with its own generous minutes allowance — you talk to Odell, Odell talks back. Onboarding interviews are voice-first either way, because people explain things best out loud. Base Odell is typed Ask, uploads, and email-in.

Because the honest price depends on your company: how many knowledge-holders we're interviewing, how deep your trust requirements run, and which lane you choose — Managed, Lease, or Dedicated. Pricing is per business, never per seat, and it's set in a conversation. The first conversation is free.

Yes. Your cabinets, your documents, your log — it's your company's brain, and it goes with you. On the Dedicated lane it never left your server in the first place.

No — and you should run from anyone who does. Nobody honest can promise “never, ever, any risk” about any system: not us, not your bank, not the safe in your office. What we promise is the stack — the contracts, the encryption, the permissions, and the log — and that you can check all of it. Here's the full trust picture.