In Grok Bot v0.75.8, tell your Bot an Auto-review rule in plain words and confirm the card it shows. Here are the steps and the limits.
All Grok Bot questions: https://www.soundminds.ai/faq.
Last updated Oct 27, 2026
You can tell your Bot the rule you want in plain words, and it can show that rule on a card for you to confirm. I checked the Grok Bot changelog for version 0.75.8, and it says, "In the app, a Bot can offer to add, change, or remove one of your Auto-review rules on a card you confirm." This needs Grok Bot 0.75.2 or later.
I haven't tested this in my own account yet. These steps come from the changelog, xAI's docs, and posts from people on X. I didn't find anything in the changelog about the card beyond "a card you confirm."
Step 1: Open a chat with your Bot. This needs Grok Bot 0.75.2 or later, and Support, then About, shows the version you have.
Step 2: Say the rule in one plain sentence about one action. For example: "Ask first before sending any external email."
Step 3: Wait for your Bot to show a card with the rule. The changelog line I read describes it as "a card you confirm," so wait to confirm until you've read it.
Step 4: Read the card before you confirm. Check whether it says Ask first or Allow automatically, and whether the action and the scope are narrow. Don't confirm a rule you can't read clearly.
Step 5: Open Settings, then General, then Auto-review, and look for the rule in your list. I checked xAI's Settings and notifications page, and it says, "Your personal rules are saved to your account, so the same rules apply on every desktop and on iPhone."
If your Bot doesn't show a card, ask once more in one plain sentence. The changelog wording is that a Bot "can offer" a rule, so my reading is that it may not always do it, though that's not something the docs say.
If there's still no card, you can open Settings, then General, then Auto-review, and add the rule there yourself. That's the place I found for adding rules in xAI's Approvals, security, and privacy page.
I checked xAI's Approvals, security, and privacy page, and it says, "Write narrow rules around a known action and scope". Two examples I found there are "Ask first before sending any external email." and "Ask first before changing a production dashboard."
Ask first rules "always stop matching actions for you," and if an Ask first rule and an Allow automatically rule both match, "Ask first wins." An Allow automatically rule only lets an action go ahead "when the automated review does not identify another reason to stop." Broad rules, such as allowing everything in the browser, are the kind to avoid.
These are posts from people on X, not from xAI. Quotes are word for word.
Both posts match the changelog line. Neither post says what the card looks like or which kinds of rule it offers.
I checked xAI's Approvals, security, and privacy page, and it says that an approval from work that started without you, such as a routine, a trigger, or a message from another Bot, "expires after about 10 minutes; the card then shows Expired, and the action does not run." I didn't find anything in the Grok Bot docs about whether a rule you add in chat changes that, so I can't tell you it does. Two people on X posted about approval cards expiring, and I list them under Issues seen on the Grok Bot v0.75.8 page.
Auto Review is model-based, and it "should complement, not replace, least privilege and explicit approval boundaries." If your team admin enforces Auto-review, your own rules "can only make behavior stricter."
I didn't find anything in the changelog about whether the chat card can add both Ask first and Allow automatically rules, or about how the card looks. I'm leaving those out rather than guessing.
Grok Bot changelog (v0.75.8): https://x.ai/changelog/bot
In the app, a Bot can offer to add, change, or remove one of your Auto-review rules on a card you confirm.
xAI docs, Approvals, security, and privacy: https://docs.x.ai/grok-bot/approvals-security-and-privacy
Write narrow rules around a known action and scope:
Ask first rules always stop matching actions for you.
Allow automatically rules let matching actions proceed only when the automated review does not identify another reason to stop.
Auto Review is model-based and should complement, not replace, least privilege and explicit approval boundaries.
Personal Auto-review rules are saved to your account, so they apply from every desktop and the mobile app.
xAI docs, Settings and notifications: https://docs.x.ai/grok-bot/settings-and-notifications
You can add your own rules on top, but they only make behavior stricter; Ask first wins when rules conflict.
@AI_Revolution (@cdepino) on X: https://x.com/cdepino/status/2114838678979907667
Tinker (@tinkerlog) on X: https://x.com/tinkerlog/status/2114847011665072498
See how I set up Grok Bot: https://www.soundminds.ai/setup.