The Freshservice Virtual Agent now understands the text present in the “tags” and “keywords” fields of a solution article. This information is used to retrieve the most appropriate solution articles for a user query, and is also used to rank those articles in order of relevance.

Tags and keywords can improve the chances for a target article to appear for a query, if it was not appearing for the same query earlier. It can also improve the rank of an article in the search results compared to the other documents, if required.

Why should you use tags and keywords?

  1. The solution article talks about a broader topic, but also applies to sub-topics that users can refer to in their queries. For example,

    1. The article is about an issue with “Office 365” but should also be shown when a user is facing an issue with “Excel” or “Word”.

    2. The article is about “JetBrains”, but should also be shown for queries referring to “IntelliJ” or “PyCharm”.

    3. The article is about installing software on a “mobile” device, and applies to both “android” and “ios” phones. Same goes for “laptop” and “macbook” / ”windows”.

    4. The article is about “zoom” or “webex” but users can also see it when they ask a question about “video conferencing tool”.

  2. There are abbreviations used in the article but users can also mention full forms of the word in their queries.

  3. There are terms used specifically in your organization to refer to certain processes.

How should you use tags and keywords?

  • Tags

    • Should convey the theme of the article. Can be used to denote the department and/or high-level type.

    • For example – “HR/IT/Finance”, “hardware/software” etc.

  • Keywords

    • Should mention the core idea of the article

    • It should also contain the entity and intent of the article and the potential user queries it will address.

    • For example – “VPN auth”, “photoshop on macbook”, “generate payslip”, “India tax benefits”, “Bangalore office wifi”, “internet connection” etc.

The feature is available for all enterprise accounts in all regions. Right now, it only understands English language text.