Kia ora Tim,
That's a good question.
Casual observations (those without supporting photos or recordings) are not currently forwarded on to GBIF (for everyone else, that's the Global Biodiversity Information Facility,
https://www.
gbif.org). Only the wild observations with confirmed identifications ("research grade") are sent on to GBIF.
The reliability of your observations without photos/recordings can still be assessed using your other observations with photos/recordings. If you've added 50 observations of fantails with photos or recordings, all correctly identified, then when you say you saw a fantail we can be confident you got the ID right.
This means that for active users of iNat NZ, observations without photos/recordings are certainly very valuable.
The more important issue, in my mind, is whether observations were collected as part of a survey ("I saw a fantail while doing a 5-minute bird count") or casually ("look, I saw a fantail"). Observations made with surveys can be used to generate robust trends and detailed habitat maps. Casual observations are good for making species lists, maps and inferring coarse trends.
So, in answer to your question, yes! You might have made one of the only kereru obs from that area that week. So it could well be useful, especially if you're other observations demonstrate that you know what a kereru looks like.
Cheers,
Jon