Choose the right AI model

Different AI models offer different strengths in:

• response quality

• reasoning

• speed

• creativity

• cost efficiency

Choosing the right model helps balance AI performance and workspace usage.

Open the AI model settings

To configure your AI model:

1. Open your AI Agent workspace

2. Go to Settings

3. Open AI Model

Inside this section, you can manage:

• AI model selection

• creativity level

• token limits

• AI behavior settings

Choose an AI model

Inside the AI Model dropdown, select the model you want your AI Agent to use.

Different models may vary in:

• reasoning quality

• response speed

• token limits

• cost per request

Some models are optimized for:

• customer support

• fast responses

• complex reasoning

• longer conversations

Understand token limits

The Max Tokens setting controls how long AI responses can be.

Higher token limits allow:

• longer answers

• more detailed explanations

• larger context windows

Lower token limits usually:

• reduce usage

• improve response speed

• keep replies shorter

Adjust creativity

The Creativity slider controls how predictable or creative the AI responses are.

Lower creativity:

• more grounded answers

• more consistent replies

• safer support behavior

Higher creativity:

• more flexible responses

• more conversational tone

• more variation in wording

For most support workflows, lower creativity usually works best.

Use AI models with instructions

AI models work together with:

• custom instructions

• synced knowledge

• snippets

• internal articles

• website content

The model generates responses using both:

• your connected knowledge

• your configured AI behavior

Test your model selection

After changing models:

1. Open the Playground

2. Ask realistic customer questions

3. Compare response quality

4. Review tone and accuracy

Testing helps identify which model works best for your workflow.

When should you change models?

Many teams switch models when they:

• need faster replies

• want better reasoning

• support larger conversations

• optimize usage costs

• improve multilingual performance

Different workflows may require different AI behavior.

Best practices for model selection

For most support workflows:

• use lower creativity

• keep responses concise

• test real customer questions

• compare results before scaling

The best model usually depends on your:

• support volume

• business complexity

• customer expectations

Next steps

• Configure AI settings

• Configure conversation starters

• Enable Human Handoff