.png%3Falt%3Dmedia%26token%3D5eafba86-93d5-4f97-b243-b2bf0af6c24a)
.png%3Falt%3Dmedia%26token%3D5eafba86-93d5-4f97-b243-b2bf0af6c24a)
Conversational AI
Help Guides
Fundamentals
Visual builder reference
Integrations
Intents and bot skills
If you're new to conversational AI, there may be some confusion about terminology. If that's the case, we advise you to check some key definitions from this manual using the links below.
The key advantage of Activechat, compared to other conversational AI and live chat automation tools, is the ability to easily combine three different approaches to chatbot creation - AI-based natural language understanding, decision trees, and natural language generation tools powered by huge machine learning models like OpenAI's GPT-3.
This makes it possible to build chatbots that can easily switch between understanding and responding to customer questions in human-like natural language, making decisions about what to say based on customer data, and automatically generating human-like responses that are both accurate and engaging.
Every automated communication in Activechat starts with detecting a customer's intent through the Natural Language Understanding (NLU) engine. Once the intent is detected, it triggers a skill, which essentially is a decision tree and allows you to build complex flows and processes with our visual conversation builder.
The decision tree model also makes it possible to build chatbots that can handle complex customer interactions, such as those that involve multiple questions and multiple possible responses for each question. You can quickly build chatbots that can handle a wide range of customer interactions, from providing basic product information and support to handling more complex customer inquiries and even taking orders and processing payments.
Every Intent should be handled by the specific Skill
There's a direct connection between the customer's intent (something that your end-user wants to achieve, from ordering a pizza to changing a subscription plan or resolving a tech support issue) and the skill in your virtual assistant that will handle this intent.
For example, if you're a pizza delivery business, your customer's intent might be to order a pizza, and the corresponding skill in your virtual assistant might be the ability to take and process orders, automatically asking specific questions like delivery address or pizza type and putting that order into your back-end CRM or ERP.
Keep in mind that not every customer's intent is the same, and your virtual assistant's skills will need to be able to handle a wide range of customer needs. However, by focusing on the key skills that your virtual assistant will need to handle your customer's intent, you'll be able to create a more effective and efficient customer experience.
Customer service automation process
Building an advanced customer service automation takes quite a lot of time and effort. Activechat helps you achieve this goal faster and more efficiently by automating three key parts of the process:
- Analyzing existing conversational data to extract the most frequent topics and intents
- Building intents with as many training phrases as possible to guarantee that your virtual assistant understands customers' queries no matter how they are phrased
- Building complex skills (conversation flows) that can handle these intents and provide necessary integrations between a conversational interface and your actual business framework
Click the link below for an in-depth overview of this process (feel free to download it as a PDF for future reference)
Last modified 2mo ago