Core Concepts
Knowing the concepts used in HaveAI helps you use the platform more comfortably. This page explains how the assistant is prepared, where you should add information and how you follow conversations.
AI Assistant
The artificial intelligence that chats with your customers. In HaveAI, Google Gemini technology is used together with your business information, behavior instructions and conversation context. You can shape the assistant's tone, language preferences and behavior from AI Behavior and Brand Information.
Training
Preparing and updating the business information the assistant will refer to when answering. Adding a document or FAQ does not mean training the base AI model from scratch.
Knowledge Base
The whole of the business information the assistant refers to. FAQs, product and service information, free-form content, uploaded documents and crawled web content make up these sources. For accurate answers, the information must be up to date, clear and consistent with each other.
Which Information Goes Where?
| Area | Purpose and example |
|---|---|
| Brand information | Describes who your business is. Example: field of activity, the services you offer and your brand description. |
| Frequently asked questions | Gives a clear answer to a specific question. Example: โAre you open on weekends?โ and your current working hours. |
| Free-form information | For broader explanations. Example: delivery, installation, maintenance or service conditions. |
| Products and services | Introduces the options you offer. Example: model, feature, scope and current price information. |
| Documents and web content | Imports your existing information. Example: catalog, user manual or service page. |
| Behavior instructions | Determines how the assistant speaks. Example: keeping answers short or asking a clarifying question when information is missing. |
Instead of adding different answers for the same topic, update the existing information. After adding new content, re-test important questions.
RAG (Retrieval-Augmented Generation)
The approach HaveAI uses to support its answers with your business information. When a question arrives, a semantic search is run in the knowledge base; the relevant content is given to the AI as context while the answer is being prepared. The search is based on finding information that matches the meaning of the question, not just the same words.
Why It Matters
Embedding (Vector)
A representation that expresses the meaning of content with numbers. In semantic search, the numeric representation of the question is compared with the representations of the content to find relevant information. You do not have to do this technical operation manually; the system runs it during content processing.
Session
โSessionโ is used in two different meanings. A conversation session refers to the talk with the customer, while a login session refers to your access to the account through a browser or device. The sessions in the Security section relate to account logins.
It helps track the conversation context and message history of a talk with a customer. The customer, the channel used and the messages are evaluated together. You can monitor conversations from the Messages screen.
Customer
A person who contacts your assistant. When shared, information such as their name, email address or phone is saved and associated with their conversations. You can reach customer information through the conversations on the Messages screen. Details: Customers.
Channel
Where the customer reaches your assistant: the web chat box, Telegram, WhatsApp, Instagram DM, Facebook Messenger or Slack. The same assistant can work across multiple connected channels with the same business information. Available channels depend on your account's entitlements and connections.
Knowledge Gap
Lets you review cases where the assistant could not answer a question with sufficient information. By checking the questions HaveAI detects and the answers given, you can complete the gaps in your knowledge base. Details: Knowledge Gaps.
Human Handoff
The routing of a conversation from the AI to a real team member. A customer requesting a representative, the assistant being unable to answer sufficiently, or rules you define can start this process. Creating a handoff request and a representative taking over the conversation are separate stages. Prepare the waiting and routing behavior according to your team's availability. Details: Human Handoff.
Sales Opportunity (Lead)
A potential customer or request identified in a conversation that your team can follow up. It is evaluated together with the contact information and the customer's need. A sales opportunity record does not mean the sale is completed.
Appointment
A scheduling process run according to your business's appointment settings. Talking about a date or time and recording the appointment are different stages. You can follow the resulting records from the Appointments section.
Usage Entitlements
The allowances your plan provides in areas such as messages, testing, training processing and channels. You can check your current limits and consumption from the Billing section; the test allowance from the AI Test screen.
API Key and Scope
An API key is the information used to authenticate the request that connects HaveAI to another piece of software. The scope determines which operations the key can perform. Do not share your keys on public pages or in code examples.
Webhook
A method that sends an automatic notification to an address you define when a supported event occurs in your account. For example, a new message event can be used to send a notification to a connected system.
API and Webhook Access
Next Steps
Dashboard Tourโ
See the panel counterparts of the concepts
AI Trainingโ
Start building your knowledge base
Was this page helpful?