Build / Chatbots
Conversations that can do more.
Create AI chatbots that understand your business, respond through the channels your customers use, and take the next action when conversation becomes work.
A chatbot is more than a chat interface.
A production chatbot combines instructions, business context, tools, permissions, a channel, and a controlled runtime. It can answer questions, use configured capabilities, preserve conversation context, and move a request toward an outcome.
Choose a channel
Meet people where the conversation already happens.
Create a bot for the audience and channel that match your business process.
Internal
Help employees answer business questions, understand records, and complete operational tasks inside the workspace.
Web
Give customers and website visitors a public conversation experience with configured sessions, message history, and attachments.
Social
Connect conversations from supported social channels through configured provider credentials and integrations.
Phone
Support voice-based chatbot interactions through the phone channel and configured call infrastructure.
Configure the bot
Give every chatbot a clear job to do.
A chatbot should be configured around a purpose, an audience, and the actions it is allowed to take.
Identity and instructions
Set the display name, description, role, and instructions that shape how the bot communicates and reasons.
Welcome message
Tell people what the bot can help with and how to get started when a conversation begins.
Capabilities
Configure the business context, tools, execution role, and permissions the bot can use.
Channel and status
Choose the channel, connect credentials when required, and control whether the bot is active and published.
Knowledge and capabilities
Useful answers come from the right context.
A chatbot becomes useful when it has the business context and capabilities needed for its purpose. Its access depends on the configured instructions, knowledge, execution role, workspace permissions, connected credentials, and enabled tools.
Configure the boundary, not just the answer.
Do not assume a bot automatically knows every record or can access every system. Give it the context and capabilities it needs, then keep its actions within the role and permissions configured for that bot.
Conversation lifecycle
From first message to useful outcome.
Public and internal conversations follow a clear operating loop.
1. Start a session
A user opens the configured channel and starts a conversation with the bot.
2. Set the context
The bot uses its welcome message, instructions, knowledge, and available capabilities.
3. Continue the conversation
Messages, attachments, and conversation history keep the exchange connected over time.
4. Resolve or hand over
The bot responds, takes the configured action, closes the request, or requests human assistance.
Human handover
Let AI handle the routine. Let people handle the important.
Eligible Web and Social chatbot conversations can be transferred to configured human agents when the conversation needs a person.
Request
The chatbot requests a handover when the configured human-assistance criteria are satisfied.
Assign
The conversation is assigned to a configured human agent or handover pool.
Transfer or close
Human agents can take the conversation, transfer it, and close the handover when the issue is resolved.
Escalate
If a response deadline is missed, the handover can escalate to another configured agent.
Handover support is channel-specific. The current capability applies to eligible Web and Social conversations, and a human should not be promised until the handover succeeds.
Example
A support assistant that knows when to act.
“Help a customer understand why a payment is pending.”
The bot welcomes the customer, understands the question, uses its configured business context, explains the next step, and requests a human handover when the case requires assistance.
Answer
Use instructions and configured knowledge to respond clearly and consistently.
Act
Use available tools and permissions when the request requires a business action.
Remember
Keep the conversation history connected so the user does not need to repeat the context.
Escalate
Move the conversation to a configured human agent when AI should not resolve it alone.
Operate the chatbot
Build, preview, publish, and improve.
Use the chatbot workspace to manage Bots, Conversations, and Knowledge. Define the purpose, write the instructions, configure the channel and credentials, preview the experience, publish it, monitor conversations, and refine the bot as the business changes.