> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orbid.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Trust, evidence, and human judgment

> What Orbid AI can draft vs what humans must own: evidence, residuals, go/no-go, and submission.

# Trust, evidence, and human judgment

Orbid accelerates bid prep with grounded tools. **Your organization owns judgment, pricing, regulatory sign-off, and buyer submission.** Prefer empty residuals over invented values.

## Ownership split

* **Scan and structure tender files** — Confirm ambiguous legal/commercial clauses
* **Match product specs to requirements** — Lock product strategy and accept deviations
* **Draft fills and documents** — Approve content before submit
* **Flag residuals and gaps** — Fill or accept residual risk
* **Support go/no-go analysis** — Make the bid / no-bid call

## Evidence discipline

* Prefer values tied to datasheets, certificates, or catalog fields
* If the agent cannot evidence a cell → residual, not fiction
* See [Evidence & ownership](/concepts/evidence-and-ownership)

## Safety defaults

* No award guarantees
* No silent substitution of product models
* Re-scan after file replace
* Re-match after material spec corrections

## Privacy baseline

Treat tenders and catalogs as customer-controlled business data. Follow your org policy and product terms. This help center is not a DPA.

## Related

* [How the agent works](/guides/how-the-agent-works) · [Fill workflow](/guides/workflows/fill-excel) · [FAQ: ownership](/faq/privacy-and-ownership)