Building Your Knowledge Base With Kapa Insights

Your knowledge base is often one of the first places that both your team and customers will visit when they have a question about or issue with your products. If they're unable to find the answer that they are looking for, or if they are confused by the answer that they receive, they're going to come to a few different conclusions:

• Your product doesn't support a feature - even if it does

• The feature is more complex than it actually is

Adding Kapa to Your Docs

Adding Kapa to your documentation can help mitigate some negative sentiment that customers may feel if they are unable to find an answer that they are looking for inside your documentation. By ingesting your documentation, Kapa gains the ability to surface documentation that customers may not usually find - ensuring that they get the right answer even if they are unable to find the correct documentation themselves.

Documentation Quality

Documentation quality is critical for any knowledge base to be effective - without it your customers or team may find it difficult to use the documentation that you've created. Quality encompasses various aspects of your knowledge base, such as:

• Completeness: is the information documented in a clear, understandable format?

• Findability: is the information easy to find?

• Accuracy: is the information correct?

• Timeliness: is the information up to date and relevant?

• Completeness: does the information contain all required information?

Kapa can solve some of these problems but it is important to remember that having good quality documentation benefits your Kapa instance, too - consider Kapa another end user of your documentation. As Kapa makes use of your documentation in order to answer queries it's important that you are feeding it with good quality sources. This will reduce it's uncertainty within answers and save your customer support team more time as more end users get immediate answers to their questions, bypassing customer support teams.

Implementing Kapa Insights

Kapa comes built in with a few tools that give you insights on what your users are asking - this can help you identify topics that are important to your users, but also topics where your documentation may have room for improvement.

Top Questions and Coverage Gaps

Top Questions and Coverage Gaps are two features that are included as standard in your Kapa instance. These two features analyse your conversations to surface insights that can be used to improve your documentation.

Top Questions analyses your conversations to determine the topics that your users are asking Kapa about - for example, your users may be having a lot of conversations on a feature that your product offers. Coverage Gaps analyses your conversations where Kapa hasn't been able to provide a conclusive answer - for example, your users may be asking about a setting within a feature that is not documented.

If you're interested in learning more about using insights from these features you can check out our blog Your Users Are Telling You What's Missing from Your Docs - Here's How to Listen

Improving Your Docs

Understanding what's important to your customers, and what's missing from your documentation, is a great method of improving your knowledge base. Remember, knowledge bases are living, waiting to be updated when new features are launched, or improved when new insights are available.

Let's look at an insight our own instance of Kapa has surfaced:

Extend the web-crawling docs with an “Editing an existing source” guide, list allowable pattern syntax (wildcards vs. regex), explain UI states like greyed-out fields, and clarify if changes require re-ingestion.

This provides us with some clear direction on what needs to be updated based on 4 conversations that users had with Kapa. To action this, we would:

1. Edit the Website Crawl data source documentation to explicitly mention that editing is not available on existing sources. This would also clarify if changes require the content to be re-ingested.

2. Create a new section documenting some pattern syntax examples that can be used.

3. Create a new section documenting what disable fields mean and that they are not editable.

The next time a user asks a similar question to one that was surfaced by Coverage Gaps, Kapa should use this updated documentation to answer their question (after it's been re-ingested).

Using LLMs to Aid in Documentation Updates

Using additional LLMs to aid in documentation updates can help speed up documentation updates that you want to make based on the insights that you now have access to. We try make this easy by adding a 'Copy for LLM' to Top Questions and Coverage Gaps - this copies all the relevant information to your clipboard, including conversation titles, so you can paste it directly into an LLM. Let's look at an example:

Kapa.ai Coverage Gap: Editing crawl URLs and patterns

Finding

Documentation covers initial crawler configuration but omits whether and how to edit start URLs or include/exclude patterns post-ingestion, why fields might be greyed out, and whether regex is supported, so the bot cannot confirm any of these abilities.

Suggestion

Extend the web-crawling docs with an “Editing an existing source” guide, list allowable pattern syntax (wildcards vs. regex), explain UI states like greyed-out fields, and clarify if changes require re-ingestion.

Conversations

• can i edit a source/website crawl to edit the start URLs after it has been ingested?

• Is there a way to add urls to exclude after a source has been configured the first time?

• Why are my URLs to exclude on my site crawl greyed out?

• can i use a regex for include urls to be crawled

–––
Explore user conversations on:
<link to cluster>

This makes it super easy to make updates, using real information from your conversations, to improve your documentation.

Conclusion

Improving your knowledge base has many benefits that can lead to your end users being more informed about your product and it's features. The implementation of insights from your Kapa conversations can help you know what needs to be added, or improved - this feedback cycle can help ensure that your knowledge base quality meets the needs of your end users.