Search result on Rovo are sometimes a mystery to me :-)

[Dikla Haimpur]

I am the head of innovation in a consulting company and am eager to understand ROVO in the best way possible.

PArt of it is gaining some hands on expericence, and whiel the courses are great, I am getting some pretty wierd answers when I actually use the tool.

I am using a site that has 0 connectors configured, and tried to ask some randon questions like: How many charcters appear on Seinfeld's second season? Geetting an answer (24 characters).

When I ask for the data source of the answer, I get a Confluence word document that has nothing to do with Seinfeld...

Same goes for other random questions - This makes it hard to trust Rovo...

[Dr Valeri Colon _Connect Centric_ · Accepted solution]

Hi Dikla! Great to see your interest in Rovo and that you're diving in hands-on.

Without data to pull from, AI can fall back on general patterns—which sometimes means confident-sounding answers with unrelated sources (like that random Confluence doc).

It’s worth trying again with some content —like a sample Confluence space—to see how it handles more grounded questions. 

Appreciate you sharing this—getting real-world feedback like yours is what helps everyone understand Rovo better.

[Ajay _view26_]

@Dr Valeri Colon _Connect Centric_  Perhaps all it takes is to check the context retrieval and handle it :)