Knowledge Base Answers
The AI retrieves answers from your uploaded documents, help centre articles, and FAQs — ensuring every response is accurate and on-brand.
AI Agent Chat Handoff
Workflow Automation
Hand off the conversation to your trained AI agent. Using retrieval-augmented generation (RAG), the AI answers questions directly from your knowledge base — handling FAQs, troubleshooting issues, and resolving queries without human intervention. When it can't help, it seamlessly escalates back to the workflow or to a live agent.
- No credit card
- 14-day free trial
- UK hosted
- Visual builder
What Is the AI Agent Chat Handoff?
The AI Agent Chat Handoff module transfers the conversation from your automated workflow to a fully trained AI agent. Unlike static chatbot scripts that follow rigid decision trees, this AI agent uses retrieval-augmented generation (RAG) to pull answers directly from your uploaded knowledge base — product documentation, help articles, policy pages, and any other content you've provided. The result is a conversational experience that feels natural and is grounded in your actual business information.
When a visitor asks a question, the AI searches your knowledge base in real time, composes a relevant answer, and delivers it within the chat window. If the visitor's query falls outside the AI's confidence threshold, or if the visitor explicitly requests a human, the module can trigger an automatic escalation — routing back into the workflow, creating a support ticket, or connecting to a live agent. This hybrid approach means your team only handles the conversations that genuinely need a human touch.
24/7 Availability
Your AI agent never sleeps. It handles visitor questions around the clock, even when your human team is offline or at capacity.
Smart Escalation
When the AI detects low confidence or complex requests, it automatically escalates to a live agent — with full conversation context attached.
Deflection Analytics
Track how many queries the AI resolves without human help. Identify knowledge gaps and continuously improve your content library.
Why Use the AI Agent Chat Handoff?
Support teams face a fundamental scaling problem: every new customer adds to the queue, but hiring agents is slow and expensive. The AI Agent Chat Handoff solves this by placing an intelligent layer between your visitors and your human team. The AI handles the repetitive, well-documented questions — "What are your opening hours?", "How do I reset my password?", "What's your returns policy?" — while your agents focus on complex, high-value conversations.
Because the AI is powered by your own knowledge base rather than a generic language model, the answers it gives are specific to your business. There's no hallucination risk from unrelated training data — the AI only references documents you've explicitly uploaded. And when it encounters something it can't answer, it doesn't guess. It escalates gracefully, passing the full transcript to your human agent so the visitor never has to repeat themselves.
Businesses using AI chat handoff typically see a 40–60% reduction in live agent volume within the first month. That's not just a cost saving — it's faster response times for the conversations that do reach a human, higher visitor satisfaction, and less burnout for your support team.
When to Use the AI Agent Chat Handoff
The AI Agent Chat module is most effective when you have a well-documented product or service and a significant portion of your support queries are repetitive. Here are the most common deployment scenarios:
- FAQ deflection: Let the AI handle common questions about pricing, policies, opening hours, and product features.
- After-hours support: Provide meaningful help when your human team is offline, instead of just showing an offline form.
- First-line triage: Place the AI at the start of your workflow to attempt resolution before involving a human agent.
- Onboarding & product guidance: Walk new users through setup steps, feature explanations, and getting-started guides.
- Internal helpdesks: Deploy the AI on your intranet to answer employee questions about HR policies, IT procedures, and company benefits.
Example: SaaS Product Support with AI Triage
A software company uses the AI Agent Chat to handle first-line support, only escalating to a human when the AI can't resolve the issue.
AI-First Support Workflow
The AI answers product questions from the knowledge base, then routes unresolved issues to a live agent with full context.
START
AI Agent Chat
Answers from product docs
Resolved
Unresolved
Human Chat
Senior agent
Step-by-Step Breakdown
- Welcome Message: A greeting message sets expectations and lets the visitor know they're chatting with an AI assistant powered by the company's documentation.
- AI Agent Chat: The AI searches the product knowledge base in real time using RAG. It composes a contextual answer and delivers it conversationally. The visitor can ask follow-up questions.
- Resolution Check: A conditional node evaluates whether the AI successfully answered the question — based on visitor confirmation or confidence score.
- Resolved Path: If resolved, the visitor sees a brief feedback prompt to rate the experience, helping you measure AI effectiveness.
- Escalation Path: If unresolved, the conversation transfers to a senior support agent with the full transcript and AI analysis attached.
Best Practices & Tips
Maximise the value of the AI Agent Chat module with these proven strategies from teams running high-volume AI support workflows.
Curate Your Knowledge Base
The AI is only as good as the content it draws from. Keep your knowledge base current, well-structured, and free of contradictions. Short, focused articles perform better than long, rambling pages.
Always Offer an Escalation Path
Never trap visitors in an AI-only loop. Always provide a clear way to reach a human agent. Visitors who know they can escalate are more patient with the AI and more likely to accept its answers.
Set Expectations Early
Use a Message node before the AI handoff to tell visitors they're chatting with an AI assistant. Transparency builds trust and reduces frustration when the AI doesn't have a perfect answer.
Review Unanswered Queries
Regularly check the queries the AI couldn't resolve. These are gold — they tell you exactly what content to add to your knowledge base to improve deflection rates over time.
Frequently Asked Questions
How does the AI Agent get its answers?
The AI uses retrieval-augmented generation (RAG). When a visitor asks a question, the system searches your uploaded knowledge base documents for relevant content, then composes a natural-language answer grounded in that content. It does not generate answers from generic training data — only from your specific materials.
What happens when the AI can't answer a question?
When the AI's confidence falls below a configurable threshold, it can automatically escalate to a live agent, route back into the workflow for alternative handling, or present the visitor with an option to speak to a human. The full conversation transcript is passed along so no context is lost.
Can I control the AI agent's tone and personality?
Yes. You can configure a system prompt that defines the AI's personality, tone, and boundaries. Tell it to be formal or casual, to always greet visitors by name, or to avoid certain topics. The AI follows your instructions while still drawing answers from your knowledge base.
Does the AI Agent Chat work outside business hours?
Absolutely. The AI agent operates 24/7 regardless of your team's availability. You can combine it with the Time of Day module to run AI-only support after hours and hybrid AI + human support during business hours.
Related Modules
The AI Agent Chat pairs naturally with these modules to build complete intelligent support workflows.
Automate Support with AI Chat Today
Upload your knowledge base, connect the AI Agent Chat module, and start resolving visitor queries automatically — no code required.