Help center federated search lets your end users see content in your help center search results that is external to your help center. This means that when an end user searches in your help center, the search results are not just limited to your help center articles or community posts. Federated search can extend its search to, for example: external knowledge bases, learning management software, blogs, and pages of your website. Your end users can also filter their search by type (for example by blog posts).
You can use either of the following methods to implement federated search in your help center:
- 联邦搜索API- REST API that lets you ingest records of your external content into the Zendesk search indexes. This method requires that your developers build and maintain a middleware layer to integrate the site that hosts the external content with the help center.
- Search crawler- Configurable in the Search settings in Guide, the search crawler lets you implement federated search in your help center without developer resources. You can set up multiple crawlers in your help center search settings to crawl and index different content in the same or different websites.
You can use the API and the crawler simultaneously. However, if you delete a source or a type via the API, then any crawler that is creating or updating records for the deleted source or type will stop working.
This article covers the following topics:
- 搜索爬虫是如何运作的吗
- How the Federated Search API works
- External content sources and types
- Known limitations for federated search
搜索爬虫是如何运作的吗
You can set up one or more search crawlers to crawl and index external content that you want to make available to users performing a search in your help center. External sites that you want to crawl must have asitemapthat lists the pages for the search crawler. In addition, the pages you want to crawl must be public (non-authenticated). If you want to add authenticated content to help center search, you must use theexternal content API.
When they are configured, crawlers run once every 12 to 24 hours, visiting the pages in the sitemap you specified during setup and ingesting content from those sources into the help center search indexes. Search crawlers index content that is in the page source on the initial page load, even if it's hidden by a UI element, such as an accordion. However, since crawlers do not run JavaScript, they do not crawl content that is rendered by JavaScript or other content that is dynamically rendered after the initial page load. Crawlers also don't crawl links on the pages they visit; they only visit the pages in the sitemap that they are configured to use.
You can set up multiple crawlers to crawl and index different content in the same website, or in different websites. SeeSetting up the search crawler.
How the Federated Search API works
Zendesk provides a REST API that lets your developers build custom integrations to ingest external content records into your help center search indexes. To ingest external content for search, developers must integrate the application hosting your external content with the external content API.
Once configured, the API ingests external content records and adds them to the help center search indexes. Indexed records can then be made available to the help center search engine when responding to user search queries.
For more information, seeSetting up the Zendesk Federated Search external content API.
External content sources and types
Regardless of the setup method you use, each content record ingested by federated search is associated with a content source and a content type, which are used to filter search results by your end user. SeeEdit search filters.
You can define sources and types either using the Federated Search API or when setting up the search crawler in Guide. If you define sources and types using the API, they will be available for selection during the search crawler setup. Similarly, if you create a content source and type when you set up the search crawler, they will be reflected in the external_content_type and external_content_source API. SeeZendesk Federated Search API Reference.
You can configure your search results to include or exclude external content sources. When external content is included in a help center search, search filters group content by source and type names, making it easier for users to find the information they are looking for. “Source” groups content by point of origin (for example blog) and “Type” groups content by the kind of content it is (for example developer documentation).
Consider useful groupings and names for your external content sources and types when creating them, as that will help your end users easily filter and locate the content that they are searching for in your help center. SeeEdit search filters.
Known limitations for federated search
At present, the following are all known limitations for federated search:
External content search results appear in help center searches, in theknowledge section of the context panel, and inUnified Search APIresponses. They do not appear in other Guide search-powered interfaces such as Instant Search or the Article Search API.
34 Comments
Hi, if I use Copenhagen but the theme has been duplicated and then customised, will we still need to edit the theme ourselves or will it be updated as the Copenhagen themes are to be updated by Zendesk? Thanks
Hi Kelsey,
Once you customize the theme, then you don't get the automatic theme updates from Zendesk.
Thanks,
Maggie
Hi,
I'm very excited about this new feature but under Help Center settings, there is no category 'Search Settings' where I can configure the Federated Search.
Is anyone else not seeing this option in their Help Center settings?
Thanks, Amanda
Hey@...
This feature is currently only available for Guide/Suite Enterprise customers.Check you Guide subscriptionto see if it is available.
Augusto,
According to Ray at Zendesk, in addition to having the Guide/Suite Enterprise edition, you must also have the Support Enterprise edition, correct?
Hey@...,
Just checked with the team and you don't need Support Enterprise for this specific feature.
You do need Support Enterprise toEnable search across multiple help centers
Hope this clarifies it
Wanted to see if anyone could share a sample scripting for an external content record via REST API - thanks.
I wonder if federated search results will be accessible by AnswerBot and in Agent Workspace?
Hi@...both of those features are on our roadmap and expected to be delivered in 2021. Making federated search results available in Agent Workspace is the first of the two to be delivered.
Is there a specific reason why articles need to be added through API or am I missing something? Also happy to hear that there are plans for these to be accessible by AnswerBot and Agent Workspace!
Hi@...,我们选择暴露一个API来摄取走读生al content records as the first interface because it allows for full flexibility to integrate with any other service that hosts the content you want to make discoverable through Help Center search, but we are currently working on adding no-code methods for adding external records as well. Unfortunately I can't give an ETA yet.
One thing we are noticing is that article filtering based upon title and URL is not unique enough. While we are using the description field, it would be good if federated search also supported the attribute to pull in tags (similar to how it works with Guide articles) to be able to present better search results.
Hi,
I'm looking for the best method to ingest content from an external website. I've tried testing the curl method forcreating an external content record, but I am getting errors. Each property gets this message "curl: (6) Could not resolve host:".
I have successfully added my content source and types.
Also, does anyone have a suggestion on a quick way to create records for each page of an external website?
Thanks, Laura
Hey@..., can you send us the curl request that you're trying so we can take a look?
Thanks!
Eric
Hi Eric,
Sure, I did end up opening a ticket for this issue as well.
So I tried two methods. The first was creating a text file with the JSON data and then used the following curl request in my Command Prompt. I've edited out a few things for sharing.
Where the CreateContent.txt file included the following.
The second method I tried was just using a curl request.
Please let me know if you need anything else.
Thank you,
Laura
I was able to resolve my create content records issue by using the Postman application instead.
Here is a great article for anyone else who is new to APIs with Zendesk.
https://developer.zendesk.com/documentation/developer-tools/working-with-the-zendesk-apis/exploring-zendesk-apis-with-postman/
我希望能够设置的基URLsources. For example, if my blog is atwww.[domain].com/blogI want the source link to go there. Currently, it looks like the link is autogenerated towww.[domain].comURL. I can change this via customization in the theme but it would be nice to have that as part of the create source API. Is that something you may support in the future?
we are using zendesk suite enterprise version .but the search is not visible to me in our settings menu.
could you please help.
Hi Gaurav,
You need to add External Content via API first. As soon as you add one External content Record (which will require at least one Type and Source), the option should be enabled for you.
To create external content, please visit the following resources:
I hope this helps! Let me know if you have any further questions.
HiGorka Cardona-Lauridsen,
in a previous post you stated:
Is Federated search live for the Agent Workspace today (Specifically in 'Knowledge Capture')? If not, do we know then the live date is?
We have 'suggested articles' live on our Support Ticket Submission flow (as a means for ticket deflection) -- if we implement Federated Search, will external articles generate in the those results?
Thanks in advance!
HelloGorka Cardona-Lauridsenand rest of the Product Management team,
Please let me know if federated search is now available for the articles/content that Answer Bot recommends. Curious also about the availability of that functionality for the Agent Workspace search as well.
Thanks.
HiWhatsApp Connector, great question!
Unfortunately Federated search is not yet available in Answer Bot answers, but we intend to make it available to Answer Bot as well. I can't currently give an ETA though.
The good news on the other hand is that Federated search (external content) is available in the Knowledge tab in the Agent Workspace. See:
If we are using external articles, how do we add labels so that Answer Bot will pick them up?
Thanks.
I double checked on my end and it doesn't look like this is possible to link up Answerbot with your external articles. The Answerbot API can only work with Help Center content in Guide.
Let me know if you have any other questions!
Hello! Will external content show up in the customer side of the ticket request form? That would be massively helpful for us (as well as appearing in the agent workspace, which it looks like it will). Thank you!
This Federated Search in your Help Center is limited to the search functionality at the moment. But this could be a great feature while making a request in the ticket request form. We encourage you to post this in ourguide product feedbackfor our product team's reference. Lastly, you can consult our professional services team as they can make custom apps outside the native feature. Thank you!
Hi, it would be great to associate the external content with existing zd categories so that when you do a search 'By Category' the external content will show up in under that category's search results.
I can try to fudge it by using the associated external record's 'Type' or 'Source'. i.e. create a 'Type' record with the same name as the category i.e. a 'Widget A' category would have a 'Widget A' 'type' record. Not a great solution though.
Is it possible to include Vimeo videos in the search results, so you could return videos hosted externally instead of having to embed them into articles?
HiKel S.,
Yes, it is possible. You just need to have an appropriate description in the Federated Search index for each video.
Hi All
I'm guessing that external knowledge (federated search) is still not available for AnswerBot or Flowbuilder?
Just wondering if this is still on the roadmap? No stress on ETA, just want to know if it's being pursued.
Cheers :)
Pleasesign into leave a comment.