Skip to main content
Brand voice training is how you turn a generic AI agent into a rep that genuinely sounds like your business. You walk through a click-through setup that teaches the agent your products, how your brand speaks, the objections your customers raise most often, and any rules it should always follow. There’s no code to write and no prompt engineering involved — just fill in what you know about your brand, and the agent learns from it immediately.

​
What you configure

Each part of the brand voice setup is independent — you can fill them in any order and update them at any time.
List the products or services your agent should know about. For each one, include the name, price, key features, and any active offers or promotions. The more detail you provide here, the more accurately the agent can answer product questions and recommend the right item to the right buyer.
Describe how your brand speaks. Is it casual and friendly, polished and professional, high-energy, or warm and supportive? A short description works well — but if you have examples of replies your team has written that you’re proud of, paste those in. The agent will match that voice immediately.
Add the objections your customers raise most often — price, shipping time, sizing, returns, or anything specific to your products. For each objection, write the ideal response you’d want a team member to give. The agent uses these as its playbook when it detects hesitation or pushback.
Set hard rules for things the agent should always or never say. Examples: always mention free shipping on orders over $50, never discuss competitor products, always end DMs with a direct link to the product page. Rules override everything else — they’re the non-negotiables.

​
Training your agent

1

Open your agent

From the dashboard, find your agent and click Edit to open the brand voice settings.
2

Add your products

Fill in the Products section with your offer details — names, prices, features, and any current promotions. Add every product your agent might be asked about.
3

Set your tone

Write a short description of how your brand communicates, or paste in 2–3 examples of replies that already sound exactly right. Either approach works.
4

Add your objections

List your top 3–5 most common objections and write the best response to each. Focus on the objections that come up most in your comments and DMs.
5

Preview the output

Click Save and then Preview to test a sample conversation. Type in a question or objection and see how the agent responds. Adjust your inputs if anything sounds off.
6

Activate

When the preview looks and sounds like you, click Activate. Your agent goes live immediately with the voice you’ve trained in.

​
Continuous improvement

Your agent doesn’t stay static after you activate it. Every real conversation it handles — and every outcome (a sale closed, an escalation triggered, a message left on read) — feeds back into the agent automatically. Over time, replies get sharper, objection handling improves, and the agent learns which approaches actually convert for your specific audience. You don’t need to retrain it manually; improvement is built into how Exerta works.
Paste 2–3 examples of your best human replies to common questions in the Tone section. The agent will match that voice immediately — no lengthy description required.

Related topics

AI Agents Overview: How Exerta Automates EngagementGet Started with Exerta: Live AI Agents in 5 MinutesWhat Is Exerta? AI Engagement Agents for Social AdsExerta Pricing: Plans, Credits, and Billing DetailsComment Moderation: Protect Your Ads Automatically