Most businesses get the same questions over and over: what's included, how does it work, which option is right for me? Answering by hand works at low volume and becomes a bottleneck at scale. This playbook sets up your AI Agent as a knowledgeable advisor that handles those conversations for you, any time of day.

An AI Agent that knows your business, answers questions accurately and in your voice, and guides contacts toward the right next step. It handles follow-ups naturally and hands off to a human only when it hits something it genuinely can't answer.

The advisor is only as good as what it knows, so start here. In AI Agent → Knowledge access and behaviour, add everything a contact might ask about: your services and pricing, what's included, your process, the FAQs you answer by hand today, and your policies on refunds, cancellations, and availability.

Then set the knowledge-access mode to Limit to instructions. This keeps the agent to what you've taught it and your stated goals, so it won't invent pricing or guess a policy. If it doesn't know something, it says so. For the full setup, including the website and document import options, see Knowledge Base: Teaching Your AI Agent What to Know.

In AI Agent → Voice and personality, pick a role that fits an advisor. Support agent suits pure question-answering; Lead nurturer suits an advisor who should also spot and warm up potential buyers.

Aim for knowledgeable, calm, and helpful without being pushy. Add a few real message examples from your inbox so the replies sound like you. The personality controls are covered in full in Configuring Your AI Agent's Personality and Tone.

In AI Agent → Goals and instructions, use Goals and capabilities to set the agent's main objective: answer incoming questions accurately and guide contacts toward the right option or next step. This keeps it focused on helping people decide rather than pushing a sale.

In Account description, explain what your business does and the kinds of decisions your contacts are weighing. Use Other instructions for the edges: what to do when it doesn't know something ("acknowledge the gap and offer to connect them with the team"), and any topics it should never speculate on, like unlisted pricing or timelines. See Goals and Instructions: Telling Your AI Agent What to Do.

In AI Agent → Activation rules, choose All pending messages for a pure advisor, so it answers any incoming DM that doesn't already have a scenario running. If you only want it to step in on clear questions, use Specific intents only and define intents like "asks about the service" or "wants to understand the options". See Activation Rules: Controlling When Your AI Agent Replies.

In AI Agent → Human intervention, turn on detection so the agent recognises when a contact asks for a real person. Add a folder to flag those contacts and a notification so your team knows to step in. See Human Intervention: Handing Off to a Human for the setup.

A contact DMs: "What's the difference between your starter and pro packages?"

The agent pulls the answer from your knowledge base, explains both options plainly, and asks one question to understand what the contact needs. It follows up naturally, and when the contact asks something outside its knowledge, it says so clearly and offers to connect them with your team, which routes them to your flagged folder.

Advisor that qualifies. Switch the role to Lead nurturer and add a couple of properties to User Data Collection: Saving Contact Info Automatically so the agent captures budget or timeline while it advises.

Advisor inside a funnel. Run a Comment-to-DM flow to open conversations, then end the flow with Hand over to AI Agent so the advisor takes over from there.

A well-fed advisor handles the bulk of your routine questions on its own, which frees your team for the conversations that actually need a person. Expect the share of DMs your team touches to drop noticeably within the first couple of weeks as you close knowledge gaps.

Once your advisor is live, tighten how it performs over time with AI Agent Best Practices, and make sure it hands off cleanly with Human Intervention: Handing Off to a Human.