Zendesk Sunshine Integration: AI + Live Agent Setup (2026)
Sandra Dajic
5 min read

Customer support. Customer support. Customer support.
Take care of your customers.
Sorry to start this article like this, but sometimes we forget how important customer support really is. A product isn’t just about features. A product is customer support.
![[object Object]](/_next/image?url=https%3A%2F%2Fcdn.sanity.io%2Fimages%2Fi6kpkyc7%2Fprod-dataset%2Fdc0300f83ba0ed7f36b9a1fe3980c0cf7c4862e8-602x401.jpg&w=3840&q=75)
Now, let’s get to the actual topic: Zendesk Sunshine integration.
Most AI support tools force a binary choice. Either the AI handles everything, or every conversation gets routed to a human agent. The real challenge is what happens in between: the handoff itself. A customer explains their problem to an AI agent, gets transferred to a human, and then has to repeat everything from the beginning. That experience erodes trust faster than a slow response time ever could.
Zendesk Sunshine Conversations solves the messaging side of this equation. It unifies customer conversations across channels into a single thread that human agents can pick up in real time. Chatbase solves the AI side. It deploys an intelligent agent trained on your company's knowledge base that resolves the majority of incoming questions without human involvement.
The Chatbase and Zendesk Sunshine integration connects both sides. When Chatbase's AI agent encounters a conversation it cannot resolve with full confidence, it transfers that conversation to a live agent through Zendesk Sunshine with the entire context attached. No repeated information. No blind handoffs. No gaps.
This guide walks through exactly how the integration works, how to set it up, and when your AI agent should escalate to a human. Whether you run a five-person support team or a global operation, the logic applies the same way.
What Is Zendesk Sunshine Conversations?
Zendesk Sunshine Conversations is Zendesk's messaging infrastructure layer. While standard Zendesk revolves around a ticketing system where customers submit requests and agents respond asynchronously, Sunshine operates on a different model entirely. It is a real-time messaging platform that connects conversations across web chat, WhatsApp, Facebook Messenger, Instagram, SMS, and other channels into a single unified thread.
Think of standard Zendesk as email-style support. A ticket opens, gets assigned, gets worked, and gets closed. Sunshine Conversations is closer to a live chat experience, but with the routing intelligence and agent workspace tools that enterprise teams need. The conversation stays continuous. A customer can start on your website, continue on WhatsApp, and the agent sees one thread with full history.
This distinction matters for AI in customer service because the handoff between AI and human agents needs to happen inside a live conversation, not through a ticket queue. When an AI agent is resolving a question in real time over messaging, it cannot create a ticket and hope a human picks it up later. The customer expects continuity. Sunshine Conversations provides the infrastructure for that continuity.
Zendesk's own automation capabilities within Sunshine are limited. For a full breakdown of what their native bot can and cannot do, see this analysis of Zendesk chatbot pricing and capabilities. That is exactly the gap Chatbase fills. Where Zendesk's built-in bot handles basic FAQ deflection, Chatbase brings autonomous resolution powered by your full knowledge base, with intelligent escalation when the AI reaches its limits.
Over 10,000 support teams use Chatbase to resolve tickets autonomously before they ever reach a human agent. The Zendesk Sunshine integration connects that AI resolution layer to your live agent team.
Connect Chatbase to my Zendesk
How the Chatbase and Zendesk Sunshine Integration Works
The integration creates a direct pipeline between Chatbase's AI agent and your Zendesk Sunshine agent workspace. Here is the flow from the customer's perspective.
A customer initiates a conversation through any supported channel. Chatbase's AI agent receives the message first. It processes the question against your trained knowledge base, conversation history, and any custom actions you have configured. For straightforward questions (order status, product specs, account details, how-to guides), the AI resolves the conversation entirely. The customer gets an accurate answer in seconds. No human involvement needed.
When the AI agent determines it should not handle the conversation alone, it triggers an escalation to Zendesk Sunshine. Three types of triggers make this happen:
- Confidence threshold: The AI's confidence in its response falls below a defined threshold. Rather than deliver a potentially inaccurate answer, it routes the conversation to a human.
- Keyword and intent detection: Specific phrases or intents automatically trigger escalation. For example, any mention of "cancel my account," "speak to a manager," or "legal" can route directly to a live agent.
- Customer request: The customer explicitly asks to speak with a human. Chatbase recognizes this intent and hands off immediately.
When escalation fires, the handoff includes the full conversation transcript, customer identification details, the AI's assessment of the issue, and any relevant metadata from your integrations. The human agent in Zendesk Sunshine picks up the conversation with complete context. They see exactly what the customer asked, what the AI responded, and why escalation occurred.
From the customer's side, the transition is seamless. The conversation continues in the same thread. There is no "please hold while I transfer you" moment. The human agent simply joins the conversation already in progress.
Step by Step: Setting Up Zendesk Sunshine in Chatbase
Before you begin, confirm that you have the following in place: an active Zendesk account with Sunshine Conversations enabled (this requires a Zendesk Suite plan or a standalone Sunshine license), a Chatbase account on the Standard plan or above for advanced integration features, and admin access to both platforms.
Step 1: Create a Sunshine Conversations App
Log into your Zendesk Admin Center and navigate to the Sunshine Conversations section. Create a new app. This generates an App ID, a Key ID, and a Secret that Chatbase will use to authenticate the connection. Copy these credentials and store them securely.
Step 2: Connect Chatbase to Sunshine
In your Chatbase dashboard, open the AI agent you want to connect. Navigate to Integrations and select Zendesk Sunshine. Paste the App ID, Key ID, and Secret from Step 1. Chatbase will validate the connection and confirm that the link is active.
![[object Object]](/_next/image?url=https%3A%2F%2Fcdn.sanity.io%2Fimages%2Fi6kpkyc7%2Fprod-dataset%2F1eda5fc1f5e95fe637509a49787691f33d883fee-1916x1168.png&w=3840&q=75)
Step 3: Configure Escalation Rules
This is where the integration becomes powerful. Define the conditions under which your AI agent should hand off to a human. You can configure confidence thresholds (for example, escalate whenever the AI's confidence drops below 70%), set keyword triggers for sensitive topics, and enable the "talk to a human" intent detection so customers always have a direct path to a live agent. Start with broader escalation rules and tighten them as you monitor performance.
Step 4: Map the Handoff Data
Choose what context transfers to the human agent during escalation. At minimum, send the full conversation transcript and customer identity. For richer handoffs, include the AI's topic classification, sentiment assessment, and any customer data pulled from your CRM through Chatbase's custom actions. The more context the human agent receives, the faster they resolve the issue.
Step 5: Test the Full Flow
Run a test conversation through every escalation path. Trigger a confidence-based handoff. Trigger a keyword-based handoff. Ask to speak with a human. Verify that the conversation appears in your Zendesk Sunshine agent workspace with all expected context. Check that responses are attributed correctly and that the customer sees a continuous thread.
Teams running Chatbase alongside Zendesk Sunshine typically deflect 80% of incoming conversations with AI, routing only the remaining 20% to human agents with full context.
See how Chatbase handles the AI to human handoff
Use Cases: When AI Should Escalate to a Human
Getting escalation rules right is the difference between an AI agent that saves your team time and one that frustrates your customers. Here are the scenarios where human involvement consistently produces better outcomes.
Billing disputes and refund requests. These involve judgment calls, policy exceptions, and emotional stakes. An AI agent can pull up the billing history and present the refund policy, but the decision to make an exception requires human authority. Route these with the billing context already attached so the agent can act immediately.
Emotionally charged conversations. Complaints about service failures, product defects, or negative experiences need empathy that goes beyond scripted responses. When sentiment analysis detects frustration or anger, escalation protects both the customer relationship and your brand.
Cross-department questions. When a customer's issue spans multiple products, services, or internal teams, the AI may not have access to all the data it needs. Rather than delivering a partial answer, a clean escalation to a human who can coordinate across departments is the right move.
VIP and enterprise customers. High-value accounts with custom SLAs often expect direct access to human support. Configure escalation rules based on customer tier so that your most important accounts always reach a live agent when they need one.
Legal and compliance topics. Any conversation touching regulatory requirements, data privacy, or contractual obligations should route to a human. The risk of an AI delivering incorrect guidance in these areas is too high.
Zendesk Sunshine vs Standard Zendesk Integration in Chatbase
Chatbase offers two distinct Zendesk integrations, and they serve different purposes. Understanding the difference is critical to choosing the right setup for your team.
The Chatbase and Zendesk ticketing integration connects to Zendesk's standard ticketing system. It creates and updates tickets based on AI conversations. When a customer interacts with your Chatbase agent and the conversation requires follow-up, a ticket is created in Zendesk with the conversation details. This is asynchronous. The ticket enters a queue, gets assigned, and gets worked by an agent on their own timeline.
The Zendesk Sunshine integration operates in real time. Instead of creating a ticket that waits in a queue, it transfers the live conversation directly to a human agent who picks it up immediately. The customer stays in the same messaging thread. There is no queue, no ticket number, no "we'll get back to you."
When should you use which? If your support model is primarily email and ticket-based, the standard Zendesk integration fits. If your customers expect real-time messaging support and you need live agent handoff within the conversation, Sunshine is the right choice. Many teams use both. The AI agent resolves most conversations autonomously, creates tickets for issues that need async follow-up, and routes urgent or complex conversations to live agents through Sunshine.
Getting the Most from the Integration
Setting up the connection is the starting point. Optimizing it is where the real value compounds over time.
Train your AI agent thoroughly. The quality of your knowledge base directly determines how many conversations the AI resolves without escalation. Upload comprehensive documentation, product guides, FAQs, and policy documents. The more your AI knows, the less it needs to escalate. For practical guidance on knowledge base optimization, see how to improve AI chatbot accuracy.
Set clear escalation boundaries. Over-escalation defeats the purpose of AI automation. Under-escalation frustrates customers who need human help. Start with broader rules, then tighten them weekly based on data. If your AI is escalating 40% of conversations, something in the knowledge base needs attention. If it is escalating less than 10%, verify that it is not delivering inaccurate answers to questions it should be handing off.
Monitor conversation analytics. Chatbase provides detailed analytics on every conversation, including resolution rates, escalation reasons, and customer satisfaction. Review these weekly. Look for patterns: if the AI keeps escalating on a specific topic, add better training data for that topic. If customers frequently ask to speak with a human after receiving an AI response, investigate whether the AI's answers are falling short.
Optimize the handoff experience. Review the context that transfers during escalation. Ask your human agents whether they have enough information to help the customer immediately, or whether they still need to ask qualifying questions. Every qualifying question an agent has to ask after a handoff is a signal that the AI should be passing more context.
The teams that extract the most value from AI customer support treat the boundary between AI and human as a living system. It should shift and improve every week based on what the data shows.
Conclusion
The Zendesk Sunshine integration turns Chatbase from a standalone AI agent into one half of a complete support system. The AI handles volume. Your human agents handle complexity. Sunshine Conversations bridges the two with real-time handoffs that preserve every piece of context.
For support teams already running Zendesk, this integration means you do not have to choose between AI speed and human judgment. Deploy Chatbase to handle the 80% of conversations that follow predictable patterns. Route the remaining 20% to your live agents through Sunshine with full conversation history, customer data, and the AI's own assessment of the issue.
The result: faster resolution times, lower cost per conversation, and a customer experience that never forces anyone to repeat themselves.
Stop choosing between AI speed and human empathy. Chatbase with Zendesk Sunshine gives your customers both, and your team gets to focus on the conversations that actually need them.
Share this article:
Sandra Đajic is the Senior Marketing and Growth Lead at Chatbase. She studied Digital Communication and Media and previously served as Chief Marketing Officer at Klu. She co-hosts The Morning Maker Show, covering SaaS, AI, and side projects. At Chatbase, she writes about AI agent implementation, customer support automation, and growth strategies for businesses adopting conversational AI.







