Put the AI assistant to work — with you in charge
Ask questions in plain language, get suggestions with evidence, and set exactly how much the AI may do — from “manual only” to “auto-apply safe changes”.
What you’ll achieve
An assistant your team actually uses because they trust its boundaries: helpful on questions and drafts, advisory on judgment, and never making decisions on its own.
Why this matters
AI features fail two ways: unused because nobody trusts them, or over-trusted until they burn someone. Both are boundary failures. Set the boundaries explicitly and the assistant becomes what it should be — leverage.
The steps
- 1
Meet the assistant
The “SalesSynq AI Assistant” floats in the corner of your dashboard. Ask it about your pipeline in plain language, or use quick actions like “Find quiet deals”, “What changed this week?”, or “Create follow-up plan”.
- 2
Know what you are talking to
The banner says it up front: you are interacting with an AI assistant, and outputs are advisory and may be inaccurate. The full disclosure — what it does, what is logged, human review and appeal — is one click away under “Learn more”.
- 3
Set the operating mode
The workspace setup hub gives you the dial: “Review before apply” (everything waits for a human), “Auto-apply safe” (routine, reversible changes apply; sensitive ones wait), or “Manual only” to keep AI actions out of the way entirely.
- Approvals are explicit: changes that affect security or shared data always ask first.
- Under Admin → Data & intelligence you control the deeper policies — AI privacy rules and how warnings and AI interact.
- 4
Hold it to the evidence standard
Suggestions cite the activity they are based on, like everything else in the product. If a suggestion has no evidence you can check, treat it as a draft — the judgment stays with you.
You’re done when
- The operating mode matches your risk appetite — deliberately, not by default.
- Your team knows the assistant is advisory, and where the disclosure lives.
Do it in your own workspace
Use this guide in an approved private-beta workspace after connecting a supported source; availability and setup depend on your data and configuration.
Private beta access is invite-only and requires approval.

