The AI in K3X writes the way you tell it to. Settings → AI Behavior sets the rules for the whole workspace. Settings → Personal AI preferences holds what the AI has learned about you personally from your chats and edits.
Open Settings (gear icon), then AI Behavior under ORGANIZATION. Owners and administrators can edit it. The settings apply to every draft, agent message and Command Center reply in the workspace.
Preferred tone: for example Direct.
Message length: for example Moderate.
Structure preference: for example Hybrid (short paragraphs with the occasional list).
What the AI should say or emphasize: your value proposition, proof points, the call to action you want.
What the AI should avoid: topics, promises or claims it must not make.
Words and phrases to exclude: a list of banned words. Add the ones your team is tired of seeing in sales email.
Communication rules: toggles for No fluff, Ask one question, Allow emojis and Always include a CTA.
Tone: Direct. Length: Moderate. Structure: Hybrid.
Emphasize: the one outcome you deliver and a single next step (a 15-minute call).
Avoid: pricing in a first email, guarantees, anything about competitors.
Exclude: "I hope this finds you well", "touching base", "circle back", "synergy".
Rules: No fluff on, Ask one question on, Allow emojis off, Always include a CTA on.
Open Settings → Personal AI preferences under WORKSPACE. The page says: "Personal to you. Learned from your assistant chats, message edits, and activity. Used when the AI works on your behalf." It has three categories:
Communication style: how you like messages to sound.
User preferences: things you have asked for, such as sign-offs or formatting.
Workflow patterns: how you tend to work, such as which stage you move deals to after a demo.
You do not have to fill this in. When you edit a draft before sending, or tell the Command Center "shorter, no exclamation marks", the AI notes it here and applies it next time. Review the page now and then and remove anything that no longer fits.
AI Behavior sets the floor for everyone. Personal AI preferences adjust on top of it for the person the AI is working for. A workspace rule such as an excluded phrase always wins over a personal habit.
Drafts from the Command Center.
Emails and texts written by agents.
Replies and summaries in the Command Center chat.
After changing AI Behavior, ask the Command Center to "draft a follow-up to the person I'm on" and read it. Adjust the settings until the draft sounds like your team on a good day.
Things you can ask: questions, updates, drafts, searches
What the Command Center is and how to open it
Roles and permissions: Owner, Administrator, Sales Manager, Sales Representative
