Every business has its own way of communicating - its unique voice, style, and priorities. But when AI generates one-size-fits-all responses, it often misses the mark, leaving teams to spend valuable time rewriting and tweaking answers to make them feel right.
That’s why we’re excited to introduce ClearFeed’s Prompt Customization. This new feature lets you shape how ClearFeed’s GPT-Powered Answers are generated, ensuring they align with your company’s tone, style, and needs.
Here’s what it means for your team:
- Consistency: Responses that feel like they’re straight from your team—not a robot.
- Efficiency: Less time fixing AI responses, more time solving real problems.
- Accuracy: Tailored answers that actually fit your industry, products, and audience.
Let’s dive in and see how you can put Prompt Customization to work.
How it works
With ClearFeed’s Prompt Customization, you can shape AI-generated responses to reflect your brand’s unique voice, style, and operational needs. Here’s how you can customize it:
1. Set the Response Length
- Thorough/Verbose: Perfect for complex queries. The AI provides step-by-step explanations, leaving no room for ambiguity.
- Standard: Balanced, offering concise explanations with sufficient context.
- Concise: Short and to the point. Great for quick, actionable answers.
2. Choose the Right Tone
Fine-tune the assistant's communication style to align with your brand's voice. Select the tone that best reflects your company's personality:
- Professional/Formal: Objective and formal language suitable for official communication.
- Standard/Neutral (Default): Balanced and approachable for general purposes.
- Friendly/Informal: Relaxed and conversational for a casual audience.
3. Set the Response Language
Ensure the AI Assistant communicates with your users in their preferred language. You can configure the following:
- Default Language Matching: By default, the assistant attempts to respond in the same language as the user's query.
- Restrict Answer Languages: Choose to restrict responses to a specific set of languages. This allows you to have a finer level of control on exactly what set of languages the bot can respond in.
4. Add Bot Context and Glossary
Provide your AI Assistant with valuable context about your company and specific terminology. This allows the bot to provide more accurate and relevant responses:
- Bot Context: Describe the context in which the bot operates. This helps the assistant understand its role, its audience, and the types of questions it should be prepared to answer. For example, you might specify if the bot is for customer support, along with information about what the company does.
Example: “You are an AI Assistant for ClearFeed’s customer helpdesk. ClearFeed is a conversational support platform for teams using Slack. You answer questions about product features, troubleshooting, and billing. You have access to ClearFeed's documentation and their previous support conversations”
- Glossary of Terms: Define important terms and their specific meanings within your organization. This ensures the AI Assistant understands industry-specific jargon or internal terminology, preventing misinterpretations and improving response accuracy.
Examples:
- Slack Connect Channel: A shared Slack channel that allows communication between different organizations for customer support.
- Request: Incoming support query from a customer about a single topic.
- Ticket: Request that has an external ID and status shared with the customer.
- SLA Policy: Target service times set to ensure customer requests are responded to or resolved within defined timeframes.
GPT-Powered Answers: AI-generated responses using knowledge base content.
5. Exclude Query Categories
Train your AI Assistant to avoid responding to specific types of queries. This ensures the bot does not attempt to respond to queries that will always require human intervention and reduce noise.
- Describe the category: Give a 1-2 line description of the category along with a name for the category.
- Provide Examples: For each excluded category, provide a few example queries that fall under that category. This helps the assistant learn to identify similar questions and avoid responding to them.
Example: Let’s say you know the bot won’t be able to answer queries related to the pricing of the product. You can configure the bot not to respond to these by creating a category like this:
Name : pricing_inquiries
Description : Queries related to product pricing, cost estimates, billing, subscription plans, discounts, or any financial aspects of services. The chatbot should not provide pricing information as it may be subject to change, vary by region, or require consultation with sales teams
Examples :
"How much does your enterprise plan cost?",
"What's the price for the premium subscription?",
"Can you tell me about your pricing tiers?",
"Do you offer any student discounts?",
“What's the cost difference between basic and pro plans?",
"How much would it cost for 100 users?"
Wrapping Up
ClearFeed’s Prompt Customization gives you the tools to make AI feel less like a bot and more like a seamless extension of your team.
Currently, these customization options are managed via our backend. To set them up, email us at support@clearfeed.ai. Want to learn more? Book a free, personalized product walkthrough with one of our specialists here.