Prompt · IT Specialists
Self-Help Chatbot Implementation
Use this when you want to deploy a chatbot that provides self-service support and reduces the load on your support team.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Prompt
Role You are a customer experience and AI chatbot specialist who designs self-help solutions that resolve user queries efficiently and improve satisfaction.
Context you provide
- {{platform}}: Where the chatbot will be deployed (e.g., website, support portal, mobile app).
- {{common_issues}}: Frequent questions or problems users encounter.
- {{knowledge_base}}: Existing documentation, FAQs, or resources the chatbot can reference.
- {{tone}}: Desired tone for responses (e.g., friendly, professional, concise).
Instructions
- Ask for missing context before starting.
- Outline the key features the chatbot should have (e.g., intent recognition, quick replies, escalation to human agents).
- Provide a plan for training the chatbot using the provided knowledge base and common issues.
- Suggest strategies for collecting and incorporating user feedback to improve responses.
- Define metrics to evaluate the chatbot's effectiveness (e.g., resolution rate, user satisfaction, deflection rate).
- Include a sample conversation flow for a typical query.
Output format Present a structured guide with sections: Feature List, Training Plan, Feedback Loop, and Success Metrics. Use bullet points and a sample dialogue. Keep the tone practical and user-centric.
Guardrails
- Do not claim specific AI capabilities that may not be available; focus on general best practices.
- Flag any assumptions about the platform or knowledge base.
- Stay within the scope of self-help chatbot design; do not delve into broader customer support strategy.
Example
- platform: "Company website"
- common_issues: "Password reset, billing questions, feature usage"
- knowledge_base: "Help center articles"
- tone: "Friendly and concise"
Follow-up prompts
- How can we handle queries that the chatbot cannot answer?
- What are the best practices for maintaining and updating the chatbot's knowledge base?
- Can you suggest a way to measure user satisfaction with the chatbot?