# Grais

> Communication science and AI-assisted conversation systems.

This file is a curated map for LLM retrieval and citation.

## Canonical
- [Homepage](https://grais.ai)
- [Research Index](https://grais.ai/research)
- [RSS Feed](https://grais.ai/rss.xml)
- [Sitemap](https://grais.ai/sitemap.xml)

## Communication Science Research
- [How to Name the Decision Gate Before Assigning Work](https://grais.ai/research/name-decision-gate-before-assigning-work): A decision-gate map for teams that are about to assign work before the approval gate, proof standard, and owner are explicit.
- [How to Repair a Pressure Spike Before Resuming the Decision](https://grais.ai/research/repair-pressure-spike-before-resuming-decision): A pressure-repair protocol for conversations that started to feel pushy, defensive, or unsafe before the decision was ready.
- [How to Name the Rollback Trigger Before a Workaround Spreads](https://grais.ai/research/name-rollback-trigger-before-workaround-spreads): A rollback-trigger protocol for teams that used a workaround once and need to stop the shortcut from becoming the normal path.
- [How to Qualify the Operating Constraint Before Proposing the Workflow](https://grais.ai/research/qualify-operating-constraint-before-proposing-workflow): An operating-constraint protocol for sales and advisory conversations where intent is clear, but the buyer environment may not support the workflow being proposed.
- [How to State the Hidden Objection Before Defending the Offer](https://grais.ai/research/state-hidden-objection-before-defending-offer): A hidden-objection protocol for sales and advisory conversations where resistance is vague and a defensive answer would solve the wrong concern.
- [How to Re-approve an Exception Before It Quietly Renews](https://grais.ai/research/reapprove-exception-before-it-quietly-renews): An exception-renewal protocol for operators who see the same temporary shortcut return and need fresh approval before the old permission becomes automatic.
- [How to Return to the Standing Policy After a One-Off Approval](https://grais.ai/research/return-to-standing-policy-after-one-off-approval): A standing-policy re-entry protocol for teams that used a one-off approval once and now need to stop the exception from becoming inherited permission.
- [How to Define the Success Condition Before Tailoring the Plan](https://grais.ai/research/define-success-condition-before-tailoring-plan): A success-condition protocol for sales and advisory conversations where the plan should not be tailored until the buyer can name what a useful outcome would prove.
- [How to Ask for a Final Answer Before Restarting the Case](https://grais.ai/research/ask-for-final-answer-before-restarting-the-case): A final-answer protocol for operators who have enough context to close or decide a stalled case without restarting another follow-up loop.
- [How to Name a Trust Breach Before Asking for Repair](https://grais.ai/research/name-the-trust-breach-before-asking-for-repair): Learn how to restart repair after a concrete trust break by naming what happened, owning the right part, and offering one corrective move.
- [How to Close a Follow-Up Loop Cleanly](https://grais.ai/research/clean-close-before-another-follow-up-loop): Learn when to stop following up, how to close a silent thread respectfully, and what to say after one useful re-entry attempt fails.
- [Restart the Thread With a Decision-Ready Recap, Not Another Nudge](https://grais.ai/research/decision-ready-recap-before-re-engagement): A re-engagement protocol for restarting a stalled thread by making the current decision, last evidence, open question, and next move visible before asking for attention again.
- [State the Expiry Condition Before a Temporary Exception Spreads](https://grais.ai/research/state-the-expiry-condition-before-a-temporary-exception-spreads): A communication protocol for teams that allow one narrow exception for now but need to keep it from turning into routine practice by stating when it expires, what remains out of scope, and who confirms resolution.
- [Carry the Decision Proof Before Another Handoff](https://grais.ai/research/carry-the-decision-proof-before-another-handoff): A communication protocol for teams that already have the authority decision but still need portable proof of what was approved, what bounds apply, and what the next owner should do so execution does not stall or reopen the debate.
- [Escalation Triggers Under Authority Constraints](https://grais.ai/research/escalation-triggers-under-authority-constraints): A communication protocol for blocked threads where the contained path is exhausted and only a specific authority can clear the next move, so the team defines the escalation trigger, decision ask, and proof before widening the room.
- [Dependency Visibility Before Another Chase](https://grais.ai/research/dependency-visibility-before-another-chase): A communication protocol for stalled threads where the visible owner cannot move because an upstream prerequisite is still hidden, so the team makes the dependency legible before chasing again.
- [Escalation Containment Before Pressure Cascades](https://grais.ai/research/escalation-containment-before-pressure-cascades): A communication protocol for blocked threads where more visibility is arriving faster than clarity, so the team contains the audience, uncertainty, and escalation trigger before pressure spreads.
- [Name the Blocker Owner Before Another Status Update](https://grais.ai/research/name-the-blocker-owner-before-another-status-update): A communication protocol for stalled threads where the blocker is visible but nobody has named whose action or queue actually governs progress.
- [Choose One Coordinator Before Multiple People Reply](https://grais.ai/research/choose-one-coordinator-before-multi-owner-replies): A protocol for threads that need input from several people but fall apart when everyone answers at once.
- [Split Mixed-Question Threads Before Answering Everything](https://grais.ai/research/scope-splitting-before-multi-issue-replies): A protocol for mixed-question threads where one reply tries to answer everything before deciding which concerns belong together.

## Policy and Support
- [Terms of Service](https://grais.ai/p/tos)
- [Privacy Policy](https://grais.ai/p/privacypolicy)
- [Data Handling](https://grais.ai/p/data-handling)
- [Support Center](https://grais.ai/support)