Search your records

Search combines filters with fuzzy text matching, over your data.

How to do it

Type. Add filters. The two work together rather than competing: a filter narrows and the text matches within what is left.

Text matching is fuzzy on purpose. A name typed from a phone call, a suburb with one letter wrong, a company name recorded two different ways: all three find the record.

What proves it worked

Search for a name you know is in there, spelled slightly wrong, with a filter applied. The record comes back. That is the case worth testing, because exact-match search finds what you already knew how to find.

What search reaches, and what it does not

Your records. The ones in your tenancy.

It does not search the wider market corpus the product holds. That is not a feature we have not built yet: the corpus is not servable to a tenant, which is a decision about what we are allowed to show rather than a limit on the search box. If that changes, corpus results would arrive through the same permission rules every other read uses, and they would be subject to what your role can see.

What your role does to results

Search results are bound by the same visibility your role has everywhere else. A record you could not open does not appear in a result list, rather than appearing and refusing to open.

Speed

Search has a published budget and it is measured against a corpus at realistic size rather than against an empty test database. If search is slow for you, that is a fault and worth a ticket, not something to work around.

What it is not

It is not natural-language search. Asking a question in a sentence will match the words in the sentence. Question-shaped search that resolves to filters is a later thing, and describing it as available now would be a promise about a store that does not exist.