Estimated read time: 5 minutes, 17 seconds

In today’s fast-paced digital landscape, AI search has become an invaluable tool for finding information quickly and efficiently. At FastSpring, we aim to continually enhance our AI search capabilities, ensuring you receive accurate and relevant answers. Here, we’ll explore how to make the most of FastSpring’s Owlbot search, why your feedback is so important, and best practices for crafting effective search prompts.

Owlbot search provides numerous benefits:

  • Speed: Instant responses to your queries save you time.
  • Accuracy: AI algorithms sift through vast amounts of data to find the most relevant information.
  • Self-Service: AI assistance empowers users to find answers independently, reducing the need for support tickets.
  • Continuous Improvement: Feedback loops enable Owlbot to learn and improve over time.

Understanding AI Limitations

While AI search is a powerful tool, it’s important to understand its limitations. AI is not perfect and sometimes provides answers that include terms like “typically” because the exact information might not be in the documentation. If Owlbot provides an inaccurate answer or a “Typically…” response, it often indicates that documentation on that specific query may not exist. This is why it’s so important to provide the thumbs-up/thumbs-down feedback so we can review and make adjustments. 

Here are some examples:

  • When searching for “How to reset my password,” you might get an answer like, “Typically, you can reset your password via the account settings.”
  • For “Why is my invoice not showing up?,” the AI might respond, “Typically, invoices may not appear due to incomplete payment information.”

These responses indicate that the AI is providing the best possible answer based on available data, but that data may not cover every specific scenario.

If it seems the AI answer doesn’t solve your problem, rate it with a thumbs-down and reach out to our Support team for personal assistance.

Common Issues With Search Prompts

We often notice that search prompt problems fall into two categories:

  • Too General: One-word prompts like “invoice” or “payment” are too broad and can lead to vague or irrelevant results.
  • Too Complex: Prompts with too many variables or overly detailed questions can confuse the AI, resulting in less accurate responses.

Best Practices for Crafting Effective Search Prompts

To get the most out of AI search, follow these two basic guidelines:

1. Be Specific but Concise

Provide enough detail in your query to give the AI a clear context, without including unnecessary information that might confuse it. The goal is to strike a balance where your query is precise and focused, helping the AI deliver the most relevant answer.

Examples:

  • Too General: “Payment”
    • This is too broad and doesn’t give enough context for the AI to understand what specific information you’re looking for.
  • Too Detailed: “I need to know how my customer can update their payment method because they tried doing it last night and encountered an error message saying the card was declined and they want to switch to a different card but are unsure of the steps.”
    • This includes too much information, which can overwhelm the AI and lead to less accurate results.
  • Specific but Concise: “How can my customer update their payment method?”
    • This provides enough context to guide the AI towards the specific information you need without adding unnecessary details.

2. Avoid Ambiguity

Clearly state what you need to know. Avoid vague terms and ensure your query is straightforward, leaving no room for multiple interpretations. This helps the AI understand exactly what you are asking and provides a more accurate and relevant response.

Examples:

  • Ambiguous: “New payment feature”
    • This query is vague because it doesn’t specify which functionality you’re referring to or what information is being requested.
  • Clear and Simple: “Can I hide certain payment methods with the Preferred Payment Method?”
    • This query clearly specifies the functionality and desired action, helping the AI to provide a precise response.
Screenshot of Owlbot AI search in FastSpring documentation, gray background with white text.

By avoiding vague terms and being specific about what you need to know, it’s easier for the AI to understand your request and deliver the information you’re seeking.

Sometimes, the AI response you receive may not suffice, and you might need to delve deeper into the documentation. 

If so, you can type in your keyword and click on the Command or Control + K button in the search window (or utilize the keyboard shortcut) for a traditional document search. 

Screenshot of Owlbot AI search in FastSpring documentation, gray background with white text, with an orange arrow point to the command + K for more button.

This allows you to review the full documentation, ensuring you have all the information you need. 

Screenshot of Owlbot AI search in FastSpring documentation, gray background with white text, showing many results of links to full documentation topic pages.

You can also provide thumbs up or thumbs down feedback on each full documentation article, as well as providing feedback comments (a text box will pop up once you have clicked on one of the thumb ratings). 

Leave One-Click Feedback on the AI’s Answer

Use the thumbs-up or thumbs-down buttons to rate the AI’s response. 

This helps us understand and address any shortcomings in the way the AI created the answer, which will help us improve the AI’s answers for the next time you use it.

The Importance of Your Feedback

For Owlbot search to be truly effective, your feedback is essential. At FastSpring, we rely on thumbs-up or thumbs-down feedback to understand whether the AI’s response met your needs. 

It’s important to note that this feedback isn’t about whether you liked the answer but whether the answer was helpful and relevant.

Without your feedback, we are limited in our ability to refine and enhance our AI search capabilities. Each piece of feedback helps us fine-tune our algorithms, leading to more accurate and useful responses in the future.

FastSpring Is Your Partner for Support

By following these best practices, you can maximize the effectiveness of AI search and ensure you receive the most accurate and helpful information. 

Your feedback is not just valuable but essential in helping us improve our services. Together, we can make AI search an even more powerful tool for all FastSpring users.

Not yet benefiting from the award-winning support FastSpring provides for software companies and their customers? FastSpring handles the entire payment process, from checkout to remitting end-of-year taxes for SaaS companies. To learn more about how FastSpring can help you scale quickly, sign up for a free account or request a demo today.

Patty Novrocki
Patty Novrocki Author
Patty Novrocki is the Customer Support Knowledge Manager at FastSpring. She brings 25+ years of enhancing customer support through innovative knowledge strategies.