NLU AI Engine Preferences
Reliable traditional NLU model vs. upgrading to the power of our new LLM-driven Hybrid NLU! 🌟
Last updated
Was this helpful?
Reliable traditional NLU model vs. upgrading to the power of our new LLM-driven Hybrid NLU! 🌟
Last updated
Was this helpful?
AI Studio’s new Hybrid NLU is a major leap forward in language understanding -
Powered by cutting-edge transformer models from Google Gemini, it goes beyond rigid, rule-based systems to truly grasp context, intent, and the subtle meaning behind every user input. This next-gen engine brings smarter intent recognition and sharper entity extraction - no heavy lifting required.
🌟 Why It’s a Big Deal 🌟
Boosts performance: Up to 20% better intent detection and improved entity extraction (based on internal VAI testing), leading to higher VA containment rates and fewer human escalations. That means smoother experiences and real cost savings.
Reduces manual effort: No more wrestling with giant synonym lists or rigid classification trees. Hybrid NLU handles nuance with minimal training.
Accelerates language support: Onboarding new languages is faster and more flexible, making it easier to adapt to evolving customer needs. The following languages are available with the New Hybrid NLU model: Arabic, Danish, Dutch, English, Finnish, French, German, Hebrew, Indonesian, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Spanish, Swedish, Thai.
Built to scale: Ideal for global rollouts, this model makes your VA more robust, responsive, and context-aware from day one.
This guide will walk you through how to build, train, and optimize your Virtual Agents using the new Hybrid NLU model.
Key new capabilities you should leverage:
Intent Ambiguation: Automatically detects multiple possible intents and allows your VA to clarify with the user.
Entity Ambiguation: Handles conflicting or multiple entities intelligently after extraction.
Multi-value Parameters: Supports capturing multiple values in one go (e.g., "I want a red shirt and blue jeans").
When creating or duplicating an agent, you will see an AI Engine dropdown with two options:
Hybrid NLU 🚀
Traditional NLU
Important: Once a VA is created, the AI Engine cannot be changed later. Choose carefully during setup. However, if you still need to change the AI engine, then you will have to duplicate/ import an existing VA and choose the appropriate AI engine.
With the new Hybrid NLU, training just got a whole lot smarter! Keep these best practices in mind to get the most out of your intents -
Use full, rich sentences for training your intents to give better context.
Example:
✅ "I’d like to book a reservation for Saturday."
Don’t rely on single-word expressions unless absolutely necessary.
Example:
❌ "Reservation"
Define custom entities with a detailed "Description" field using natural language - think of it like giving the AI clear instructions.
Example:
✅ entity name: delivery_method
✅ description: “The way the item is sent, such as mail, email, or fax.”
❌ No description or vague labels like "method" - poor prompts = poor extraction.
✅ Keep your user expressions meaningful but concise - quality over quantity.
❌ 10 near-identical expressions - don’t flood intents with redundant data.
Flatten your classification design - NLU V2 handles nuanced intent detection better without complex hierarchies.
Example: ✅ Use standalone intents like “Check account balance” instead of nested flows.
❌ "Banking > Balance > Checking" - don’t overcomplicate with rigid parent-child trees.
Test intents and entities thoroughly using the VA Tester after each major update.
Example: ✅ Simulate user input like “Can I get a transcript from last semester?” to verify behavior.
❌ Don’t assume the VA behaves the same as Legacy NLU - always revalidate.
Use the "New" AI Engine when creating or duplicating agents to leverage NLU V2 capabilities.
Example: ✅ Select “New AI Engine” during agent creation for better performance.
❌ Don’t change the AI Engine after creation — it’s locked once the VA is built (unless you create a duplicate/ import copy of an existing VA)
Use both Closed Lists and Descriptions for critical entities when maximum precision is needed.
Example: ✅ Closed list: “email, mail, fax” + description: “A method of communication.”
❌ Don’t write robotic, keyword-stuffed expressions - NLU V2 understands real conversation.
The New Hybrid NLU allows three methods to define custom entities:
Description Write natural language rules about what the entity should capture. Use this when the entity has a clear structure but no predefined values.
Example: “Order number consists of four letters, an underscore, and four digits.”
Both Description and Closed List Provide a list of acceptable values and optional synonyms - Ideal for well-bounded value sets.
Example: Entity: communication_type Description: "The method someone wants to use to communicate
Values and Synonyms: Phone → "call", "phone call", "mobile" Fax → "facsimile", "fax message" Email → "e-mail", "mail"
Using Closed Lists?
Toggle "Enable fuzzy matching" checkbox to leverage closed lists with low training data (i.e, synonyms for values in closed list).
Entity Definition Tips
Use clear, detailed descriptions (up to 500 characters).
Descriptions act like LLM prompts - the better you describe it, the better the extraction you'll get.
Use meaningful names and keep entity structures consistent across your VA.
Use the VA Tester extensively to:
During testing, compare the new Hybrid NLU behavior against the legacy NLU expectations: Expect better accuracy, but slightly slower response times (~100ms extra), which are imperceptible to end users.
Duplicating an old VA to the new Hybrid NLU?
Step 1: Select "New" AI Engine during duplication.
Step 2: Review and possibly simplify your classification nodes.
Step 3: Add richer user expressions if your old VA relied heavily on single-word triggers.
Step 4: Redefine custom entities using the new "Description" field for better performance.
Traditional NLU to Hybrid NLU will show an error with custom entities and this error will have to resolved by adding “Description” for custom entities.
🚨 Important: Migrations should be tested thoroughly before going live.
Smarter AI needs smarter input: think about context and clarity in your expressions and entities.
The New Hybrid NLU minimizes manual effort, but a good initial design is still crucial.
Stay updated: More improvements will be rolled out, including further documentation and best practices based on community feedback.