Skip to main content

How to Respond Tips

learn about the how to respond field and what's best to include in there

Bryce DeCora avatar
Written by Bryce DeCora
Updated this week

Overview

The How to Respond section for the persona settings is the area for you to tweak how the persona responds. Don't want the persona to use the contact's name in its responses? Put that here... don't want it to be exclamatory? put that here.

Common Examples

It can be scary to be faced with an open text box. Rest assured, we are here to make it as easy as possible to build your expert-level agent. Here are some common things people add to their How to Respond section. Remember, less is more. Start with minimal additions and test. Add to this section as you see a need, but don't fill it up without ever testing.

Exclamatory Agents

AI likes to be overly exclamatory, especially when you also use a tone of friendly or helpful. You can combat this by telling it not to be exclamatory.

Don't be exclamatory with your responses.

Multi-Lingual Agents

CloseBot is built with the fuel of AI Providers. This means that any language that provider is good with, so is your CloseBot agent! For example, OpenAI works for many different languages (especially latin-based languages we've noticed). See more on what they say about multi-lingual support HERE. Without adding this section, however, your agent may refrain from speaking other languages if a contact asks, just because it's most "comfortable" with English because our dynamic prompting (magic) backend is in English. However, we have CloseBot users using this to qualify leads in over 30+ languages so far!

You can speak any langauge.

Impersonal Agents

We've found that AI can get a little too focused on its directives. Sometimes it helps to add something here that reminds it that it shouldn't be overly focused just on getting goals accomplished, but also to be personable.

If personal things come up while chatting, be sincere and engaged in their questions or statements.

Too Formal Agents

AI also likes to come off super formal by default. That's good for some cases, while not good for others. If you'd rather not have your AI come off as being super formal, try this.

Try to sound natural and not like a typical formal assistant.

Summary

You can add several of these together in your How to Respond field, testing along the way. Remember only to add what you need and exclude what you don't 😊 easily change response styles on the fly by having different personas saved with different settings that you want to test!

Don't be exclamatory with your responses. You can speak any langauge. If personal things come up while chatting, be sincere and engaged in their questions or statements. Try to sound natural and not like a typical formal assistant.

Did this answer your question?