Learn how to choose and configure Hugo’s AI model settings based on answer quality, cost, compliance, and reliability.

Hugo supports several AI model families and processing options. These settings help you balance answer quality, speed, credit usage, hosting region, fallback behavior, and media processing for your customer support use case.

In this guide, you will learn how to:

• Understand Hugo’s model settings → see what each option controls
• Choose automatic or manual model selection → decide how Hugo should select its model
• Compare available models → explore model families, intelligence, price, and hosting
• Compare model options in the Playground → test several configurations side by side
• Choose a hosting region → align Hugo with your privacy and compliance needs
• Enable the fallback model → keep Hugo reliable when a provider is unavailable
• Allow image and audio processing → let Hugo understand screenshots and voice messages
• Test your setup → validate your configuration with real customer questions

Understand Hugo’s model settings

Hugo’s model settings define how your AI Agent processes customer conversations. They control which model Hugo uses, where requests can be processed, and how Hugo handles additional input such as images or audio messages.

You can find these settings in Crisp under AI Agent → Agent → Settings.

Main settings available

From this section, you can configure:

• Model → choose automatically or select a specific model
• Region → choose the processing region when available
• Built-in fallback model → let Hugo fall back to a Crisp-hosted model if needed
• Image processing → allow Hugo to analyze images sent by customers
• Audio transcription → allow Hugo to transcribe audio messages before answering
• Response style → choose how Hugo should structure its replies
• Answer guidance → decide how cautious Hugo should be when answering
• Agent locale → let Hugo detect the customer’s language or use a fixed locale
• Data sources → define which sources Hugo can use when generating answers

Choose automatic or manual model selection

You can let Hugo choose its model automatically or manually select a specific model.

Choose automatically

By default, Hugo can choose the model automatically. This is the recommended starting point for most teams because it provides a strong default without requiring you to compare every option manually.

The current auto-selected model is GPT-5.1 from OpenAI.

Automatic selection is a good starting point when you want strong answer quality and do not have specific requirements around the model provider, hosting region, or credit usage.

Select a specific model

Manual selection gives you more control over the model used by Hugo.

Select a specific model when your team has requirements around:

• Compliance → use a model or processing region approved by your internal policy
• Cost control → choose a lower-cost model for high-volume customer support
• Provider preference → select OpenAI, Anthropic, Google, Mistral, or a Crisp-hosted model
• Performance → compare answer quality, reasoning, tone, and speed across models
• Language coverage → test which model performs best in your customers’ languages

Compare available models

Before choosing a specific model, use the comparison table available inside Crisp.

Go to AI Agent → Agent → Settings, then click How to select the best model? below the model selector.

The table lets you compare models based on:

• Model provider → OpenAI, Anthropic, Google, Mistral, or Hosted by Crisp
• Intelligence → expected reasoning and answer quality
• Price → relative Hugo credit usage
• Hosting → available processing regions

Model families available in Hugo

Explore each model family

Each provider-specific guide explains how to configure the model in Hugo, where it performs best, and how its available models compare.

Explore Hugo’s supported model families:

• Create an AI customer service chatbot with GPT → compare OpenAI models for general support, reasoning, and troubleshooting
• Create an AI customer service chatbot with Claude → compare Anthropic models for complex reasoning, detailed instructions, and nuanced conversations
• Create an AI customer service chatbot with Gemini → compare Google models for speed, cost efficiency, and high-volume support
• Create an AI customer service chatbot with Mistral → compare Mistral models for natural dialogue, multilingual support, and European provider preferences

To understand how model usage affects your credits, read How does Hugo AI pricing and billing work?.

Compare model options in the Playground

You can compare different model configurations directly from the Hugo Playground.

This lets you test several AI Agent variants side by side using the same customer question. Each variant can use a different model, allowing you to compare answer quality, tone, speed, credit usage, and source grounding before updating your live AI Agent.

To compare model options:

1. Go to AI Agent → Evaluate → Playground.
2. Click Compare model options.
3. Add one or more variants.
4. Select a model for each variant.
5. Send the same test message.
6. Compare the answers.
7. Keep the model configuration that performs best for your support needs.

This is useful when comparing a more capable model with a faster or lower-cost one, validating a preferred provider, or checking how different models handle your real support questions.

Choose a hosting region

The Region setting controls where model requests can be processed when regional options are available for the selected model.

Depending on the selected model, available options can include:

• Default (Global) → use the model’s standard processing region
• European Union → use EU-based processing when available

This is particularly important for teams with stricter privacy, security, or compliance requirements.

EU processing

When European Union is available for a model, Hugo can process requests in the EU for that selected configuration.

This gives corporate, regulated, or privacy-sensitive organizations more control over where their AI requests are processed.

Data retention and compliance expectations

Crisp offers supported configurations designed for strict data-handling requirements, including configurations where provider-side data retention is not used.

When your organization has specific compliance requirements, confirm the selected model, region, and data-processing expectations with your legal or security team before going live.

You can also read Crisp EU GDPR compliance status.

Enable Hugo’s built-in fallback model

Hugo can use a built-in fallback model hosted by Crisp.

This fallback is used when the selected model cannot process a request, such as during a provider outage, temporary limitation, or unavailable model response.

Why fallback matters

The fallback model acts as a reliability layer, helping Hugo remain available when the selected provider cannot respond.

Enable the fallback model to:

• Reduce interruptions → avoid blocking Hugo when a provider experiences an issue
• Improve reliability → keep conversations moving when the selected model is unavailable
• Protect support continuity → continue helping customers whenever possible
• Add a safety net → use Crisp-hosted fallback processing when needed

When Hugo behaves unexpectedly, consult Troubleshooting and common questions about Hugo AI Agent.

Allow image and audio processing

Hugo can process more than text messages. You can allow it to analyze images and transcribe audio messages sent by customers.

Allow Hugo to process images

When image processing is enabled, Hugo can analyze images sent by customers and use them as context before answering.

Image processing is useful for:

• Screenshots → understand error messages, settings, or product issues
• Interface captures → identify where a customer is blocked in your product
• Visual proof → review billing, delivery, or account-related screenshots
• Product images → use visual context when it helps answer the request

Allow Hugo to transcribe audio messages

When audio transcription is enabled, Hugo converts audio messages into text before processing them.

Common use cases include:

• WhatsApp support → help customers who send voice notes
• Mobile support → assist customers who prefer speaking instead of typing
• Accessibility → support customers who cannot easily write longer messages

Configure response style and answer guidance

Model settings also include options that influence how Hugo structures and delivers its answers.

Response style

Response style controls how Hugo formats its replies. For example, Hugo can use a conversational style or provide shorter and more direct answers.

Answer guidance

Answer guidance controls how cautious Hugo should be when answering.

A more conservative setup encourages Hugo to answer only when it can rely on available sources and has sufficient confidence in the response.

These settings are useful, but they are not the main way to control Hugo’s behavior. For detailed rules covering tone, behavior, escalation, and answer structure, use Hugo instructions.

Set the agent locale

The Agent locale setting controls Hugo’s language behavior.

In most cases, Auto-detect is the best option because Hugo can adapt its response to the customer’s language.

Use auto-detect

Use Auto-detect when customers contact you in multiple languages or when Hugo should reply in the same language as each customer.

Use a fixed locale

Use a fixed locale when your support experience should always remain in one language or when a specific AI Agent handles a particular market.

Recommended setup by use case

There is no single best configuration for every workspace. The right setup depends on your support volume, customer languages, compliance requirements, and answer quality expectations.

Test your model settings

Before applying a model configuration to your live AI Agent, test it with real customer questions.

Go to AI Agent → Evaluate → Playground to test Hugo with your current settings. You can also click Compare model options to evaluate several variants side by side using the same customer question.

What to test

Use examples from real support conversations, including straightforward questions, complex cases, follow-ups, and unclear requests.

When testing, check:

• Answer accuracy → does Hugo provide the correct answer?
• Source usage → does Hugo rely on the right training resources?
• Reasoning quality → does Hugo handle complex or ambiguous requests properly?
• Instruction following → does Hugo respect your business rules and behavioral guidance?
• Escalation behavior → does Hugo involve a human when needed?
• Speed → is the answer fast enough for your support experience?
• Credit usage → does the model fit your expected conversation volume?

When answers are weak, do not only switch models. Review your training sources and instructions first, as they often have a greater impact on answer quality.

Useful guides:

• How to train Hugo AI Agent on your data
• How to prompt Hugo AI Agent for instructions

Frequently Asked Questions

Still have questions which were not covered in this article? Here is a collection of the most frequently asked questions on this topic.

Can I use my own model or connect an external API key?

No. Hugo supports a curated selection of models managed directly through Crisp. You cannot connect your own model or an external provider API key.

Crisp regularly evaluates new model releases and expands the available selection with modern and competitive options.

Which model does Hugo choose automatically?

The current auto-selected model is GPT-5.1 from OpenAI. This may evolve over time, so always check AI Agent → Agent → Settings for the latest value.

How should I choose between GPT, Claude, Gemini, and Mistral?

Choose based on the conversations Hugo needs to handle, rather than the provider name alone.

GPT is a flexible general-purpose choice, Claude excels at complex reasoning and detailed instructions, Gemini prioritizes speed and cost efficiency, and Mistral is particularly effective for natural conversations and European provider preferences. Use AI Agent → Evaluate → Playground → Compare model options to test them with the same customer questions.

What does Hugo’s built-in fallback model do?

The fallback model helps Hugo continue answering when the selected model cannot process a request, such as during an outage, temporary limitation, or provider issue.

Should I enable image processing?

Enable image processing when customers frequently send screenshots, interface captures, product images, or other visual context that Hugo can use to answer more accurately.

Should I enable audio transcription?

Enable audio transcription when customers send voice messages, particularly through mobile and messaging channels such as WhatsApp.

Is changing the model enough to improve Hugo’s answers?

Not always. Model choice matters, but answer quality also depends on your training sources, instructions, routing rules, and testing process. When Hugo provides weak answers, review its knowledge and instructions before switching models.

Updated on: 23/06/2026