Plan requirement
| Subscription | Suite Team or higher, Knowledge Professional or higher |
| Access | Admin |
Federated search brings content from other systems into your help centre results, so people search once.
The problem
Knowledge lives in several places: your help centre, a documentation site, an intranet, a shared drive, a wiki. People have to know which to search, and they guess wrong.
Federated search puts results from all of them in one place.
What can be connected
A range of systems: file stores, documentation platforms, other Zendesk help centres, and a web crawler for anything with a public address.
The crawler is the general answer for content that has no direct connector, with two conditions attached: the pages have to be public, and you get up to 50 crawlers, each pointed either at a whole site or at up to 500 specific pages on it.
For anything behind a login, an intranet included, the crawler is not the route. That is the Federated Search API, where your own systems push records in. It works on private content precisely because you decide what goes in, and it needs a developer to build and keep it running.
Across both routes the ceiling is 50,000 external records, which is generous for documentation and small for a file store somebody points at wholesale.
Where it genuinely helps
- Internal help centres, where colleagues search four systems today.
- Technical documentation that lives outside Zendesk and will stay there.
- A second help centre whose content is relevant to the first.
What it does not do
It does not improve the other content. Results from a badly maintained wiki are badly maintained results, now surfacing next to your careful articles and undermining them.
Connect sources you would be happy for somebody to be sent to.
Permissions need thinking about
Content from another system arrives under your help centre's visibility rules rather than its own. Before connecting a source, be clear about who can see what, particularly for anything internal.
Two ways in, same requirement
There are two routes: the web crawler, which you configure in the help center search settings, and the Federated Search API, which needs a developer to build and maintain a middleware layer.
They sit on the same plan requirement, so the choice is about effort and control rather than about what you are entitled to. The crawler needs no developer at all, which for most situations settles it.
You can run several crawlers, each covering different content or a different site. Where a sitemap exists the crawler follows it; where none exists it follows links from the top page down.
Comments
0 comments
Article is closed for comments.