# Your strategy data stays with your company.

> Each organization gets its own database for OKRs and strategy. Optional AI uses private open-source models — no training on your OKR text; prompt bodies are not kept on the billing store, provider, or training path by default; current org context each run. Coach transcripts, Document Alignment Review reports, and workflow findings (when Agents are enabled) live in your company database. Uploads are not retained.

Agent index: https://gahmbit.com/llms.txt
HTML page: https://gahmbit.com/trust

Each organization gets its own database for OKRs and org strategy. Optional AI uses private open-source models — it does not train on your OKR text, does not keep prompt bodies on the billing store, provider, or training path by default, and loads current org-scoped context each run. When Agents are enabled, coach transcripts, Document Alignment Review reports, and workflow findings archive in your company database for org review — not a training corpus. Document uploads are not retained.

## Your strategy lives in your company’s database

People, cycles, OKRs, check-ins, and comments sit in a dedicated database for your organization — not a shared strategy table with other companies.

Accounts, membership, and billing live separately from strategy content. That layer holds AI run metadata (who, when, which agent, token counts, status) — not your OKR text.

After membership checks, members of org A cannot open org B’s strategy database. Isolation is between companies.

## Optional AI that does not train on your OKRs

Private open-source models behind Gahmbit’s AI service — not “your OKRs train a public chatbot.”

Prompt and completion bodies are not kept on the billing store, at the inference provider for training, or used to train models. Billing stores run metadata only — not message bodies. When Agents are enabled, coach transcripts, Document Alignment Review reports, and org-visible workflow findings are archived in your company’s isolated database so members can reopen them — product history for your organization, not a training corpus. Document uploads from Run the Gahmbit are not kept. Drive files are exported ephemerally for the run and never stored.

Each run loads current org-scoped OKRs. Continuing a coach thread can include prior turns from your company archive (those turns are billed). Org Brain, when enabled, grounds coaches in company-visible sources you uploaded — opted-in product memory, not a hidden training store.

Metering gates the run, not the read: when findings exist, every authenticated member can read them.

AI is optional prepaid leverage when enabled; the free OKR core works with a zero wallet.

## Isolation is between companies

Within an organization, radical transparency is unchanged: every authenticated member can read all OKRs; Missing coverage stays loud; there is no privacy toggle on objectives.

Optional authenticator MFA (TOTP) is free-core Auth: members can enroll, and org admins can require it — never billed as AI credits, and never a wall that hides OKRs from enrolled members.

Free core, Slack / Agents enablement, and prepaid AI pricing are unchanged by this trust story — only inference burns credits when AI is enabled.

How isolation works: One strategy database per organization — company data stays separate.
