As travelers turn to Large Language Models (LLMs) to plan their trips, they are asking highly specific, long-tail questions about your brand. However, a critical issue arises when an AI looks at your official website—your designated "Source of Truth"—and cannot find the answer.
In the Bonafide platform, these missing pieces of information are categorized as "Unverified" prompts. Unverified prompts represent the biggest content gap for most companies today because they highlight exactly what users are asking the AI, but your brand is failing to answer. When your website lacks this information, the AI is forced to guess or pull potentially inaccurate information from unverified corners of the internet.
To ensure your brand owns the narrative, your first priority should be to find these unverified questions and fill the knowledge gap. Here is exactly how to do that using the Curator module.
Step 1: Finding Your Content Gaps (Unverified Questions)
To locate the questions your website is failing to answer, you must filter your Curator dashboard to show only "Unverified" prompts.
1. Log into Curator and Access Detail View:
When you open the Curator module, click the "Detail" view icon (which looks like a grid or spreadsheet) in the upper left corner to transition away from the initial bar graphs.
2. Locate the Verified Column:
Scroll to the right side of the spreadsheet until you find the column labeled "Verified".
3. Filter for Unverified Status:
Click the dropdown arrow in the column header (which typically defaults to "Status: All"). Check the box next to "Unverified" and click OK.
Your screen will now filter the entire curation sheet to only display the questions that your brand's website currently does not answer.
Step 2: Editing and Filling the Gaps
When looking at your filtered list of Unverified questions, pay close attention to the "Official Response" column. You will encounter two distinct scenarios that require different actions.
Scenario A: The AI-Recommended Answer
In many cases, the Official Response cell will contain text, even though the row is marked as Unverified (often highlighted in yellow or gray). This means that while the AI couldn't find the answer on your official website, it found enough consensus on the open web to recommend an answer to you.
EXAMPLE
Here is how to handle these recommendations:
- Review: Read the AI's recommended answer carefully.
- Edit (If Necessary): If the AI's guess is close but not quite right, simply double-click the cell to freely edit the text, just as you would in Microsoft Excel or Google Sheets. Add context, correct errors, and delete inaccurate information.
- Verify: Once the answer perfectly reflects your brand's truth, click the "Unverified" status button to change it to "Verified" (the row will turn dark green).
By clicking "Verified," you take ownership of that answer. Even though it wasn't on your website before, it is now your brand's official response, and it will be used to train the LLMs to answer correctly in the future.
Scenario B: The Blank Answer
Sometimes, the Official Response cell will be completely blank. This indicates that all the major AI models were completely unsure of the answer; they couldn't find a definitive consensus anywhere on the web, so they refused to guess.
Here is how to fill this critical void:
- Draft the Answer: Double-click the blank Official Response cell.
- Over-Answer: Type in your comprehensive, official answer. There is no text limit, and you can easily copy and paste information from your internal documents. Because LLMs thrive on deep context, you are highly encouraged to write as much detailed information as possible rather than simple "yes" or "no" answers.
- Verify: Press Enter to save your text, and then click the "Unverified" button to change its status to "Verified".
The Result
Once you have systematically changed your unverified prompts to verified ones, you have successfully plugged the holes in your brand's digital footprint. You are effectively helping the train the AI on the exact questions it was previously failing to answer, transforming your newly curated knowledge base into the ultimate authoritative source of truth.