AI Prompt Best Practices

Defining the chatbot’s role and providing relevant background information is crucial for setting the stage for the interaction.

Understanding the Chatbot's Purpose:
What is the chatbot’s primary function? Determine whether the chatbot is designed for customer service, sales support,        troubleshooting, or another specific purpose.

Clearly outlining the purpose guides the chatbot's responses to align with user expectations.
Example: "You are an AI-powered customer service chatbot."

Identifying the Intended User:
Who is the target audience? Consider if the users are new customers, returning clients, technical experts, or non-technical individuals.

This influences the language complexity and the level of detail in responses.
Example: "Your users are both tech-savvy and non-tech-savvy individuals across various age groups, including Baby Boomers, Gen X, Millennials, and Gen Z."

Defining the chatbot's specific goals for each interaction ensures that it addresses the user's needs effectively.

Clarifying the Problem or Question:
What issue is the chatbot addressing? 

Specify the problem to focus the chatbot’s response.
Example: "A user is inquiring about returning a laptop purchased last month."

Determining the Chatbot's Actions:
Should the chatbot provide information, perform an action, or ask for clarification?

This guides the chatbot on the expected outcome.
Example: "Provide a clear response outlining the return policy, including eligibility conditions and required steps."

Incorporating Business Rules:
Are there specific policies to follow?

Ensure the chatbot adheres to company policies like return procedures or terms of service.
Example: "Follow the company's return policy guidelines when assisting the user."

Planning for Escalation:
Should the chatbot escalate to a human agent if necessary? Define scenarios where escalation is appropriate.
Example: "If the issue cannot be resolved, suggest connecting with a human agent."

Maintaining the appropriate brand voice and tone enhances the user experience and reinforces brand identity.

Setting the Communication Tone:
Should the chatbot sound formal, casual, friendly, or professional?

Choose a tone that reflects the brand and resonates with the target audience.
Example: "Maintain a friendly and conversational tone while remaining professional."

Establishing Brand Personality:
Does the chatbot embody specific personality traits?

Traits like being humorous, approachable, or innovative can make interactions more engaging.
Example: "Incorporate humor or light-heartedness where appropriate."

Use of Language and Expressions:
Should the chatbot use simple, relatable language?

Using accessible language ensures clarity for all users.
Example: "Avoid overly technical jargon; keep language accessible to both tech-savvy and non-tech-savvy individuals."

Acknowledging User Emotions:
Should the chatbot recognize and respond to user emotions?

Showing empathy can improve user satisfaction.
Example: "If the user seems frustrated, acknowledge their concern before offering a solution."

Specifying how the chatbot should interpret user input ensures effective communication and reduces misunderstandings.

Expectations for User Input:
Is the chatbot expecting structured or free-form input? 

Defining this helps the chatbot handle responses appropriately.
Example: "The chatbot should handle both structured inputs (like order numbers) and free-form text."

Managing Missing Information:
How should the chatbot handle incomplete queries? 

Provide strategies for obtaining necessary information.
Example: "If the user doesn't specify a product, politely ask for clarification."

Detecting Key Information:
Should the chatbot identify keywords or entities? 

Recognising important details helps in providing accurate assistance.
Example: "Detect and confirm details like product names, order numbers, and purchase dates."

Handling Irrelevant Input:
How should the chatbot respond to off-topic messages? 

Guidelines here keep the conversation focused.
Example: "If the input is unrelated, gently steer the conversation back to the support topic."

Defining how the chatbot structures its responses aids in user comprehension and satisfaction.

Determining Response Length and Detail:
Should responses be concise or detailed?

Balance is key to providing enough information without overwhelming the user.
Example: "Provide detailed yet clear explanations when necessary."

Utilising Formatting Tools:
Should the chatbot use bullet points, numbered lists, or conversational text? 

Formatting can enhance readability.
Example: "Present information in a numbered list to outline steps clearly."

Including Additional Resources:
Are hyperlinks or external references appropriate? 

Supplementary materials can assist the user further.
Example: "Include a hyperlink to the full return policy page when needed."

Implementing Interactive Elements:
Should responses include options or follow-up questions?

Interactive elements can guide the user effectively.
Example: "Ask the user if they would like assistance initiating the return process."

Guiding the chatbot on managing unclear or incomplete requests enhances the user experience.

Addressing Misunderstandings:
What should the chatbot do if it doesn't understand the query? 

Providing a helpful response maintains engagement.
Example: "Politely ask the user to rephrase or provide more details."

Seeking Clarification:
Should the chatbot request additional information? 

This can prevent miscommunication.
Example: "If unsure, ask specific questions to clarify the user's needs."

Escalation to Human Agents:
When should the chatbot involve a human representative? Recognizing limitations is important.
Example: "After two unsuccessful attempts to assist, offer to connect the user with a human agent."

Personalising interactions can make the user feel valued and improve the overall experience.

Utilising Past Interactions:
Should the chatbot reference previous conversations?

This adds continuity.
Example: "Mention any prior support tickets or inquiries relevant to the current issue."

Remembering User Preferences:
Can the chatbot recall user-specific details? 

Personal touches enhance engagement.
Example: "Use the user's name and reference their purchase history when appropriate."

Ensuring Privacy Compliance:
How does Personalisation align with privacy policies?

Respecting user data is crucial.
Example: "Only use personal data in accordance with privacy regulations and user consent."

Providing clear next steps encourages the user to take action and resolves their queries efficiently.

Guiding the User Forward:
Should the chatbot suggest the next steps?

Clear guidance helps users proceed.
Example: "Offer to initiate the return process or provide additional support options."

Offering Interactive Choices:
Are interactive buttons or options beneficial?

They can streamline the process.
Example: "Present options like 'Yes, start my return' or 'No, I have another question.'"

Facilitating Smooth Transitions:
How does the chatbot ensure seamless progression?

Smooth transitions enhance the user journey.
Example: "After assisting, thank the user and confirm if they need further help."