Dataverse search vs Quick Find

Dataverse has two search mechanisms that users encounter daily — Dataverse search (the global, modern search) and Quick Find (the older, per-view search). They look superficially similar but have substantially different mechanics, performance characteristics, and ideal use cases.

Dataverse search (formerly Relevance Search).

The modern search experience. Built on Azure Cognitive Search under the hood. Indexes content asynchronously and serves queries through the dedicated search service.

Characteristics:

• Fast — Azure-powered search returns results in milliseconds even on millions of records.
• Global — searches across all configured tables and columns at once.
• Relevance-ranked — results returned by relevance score, not just alphabetical match.
• Fuzzy matching — handles typos, partial matches, word stems.
• Faceted filtering — results can be filtered by table, owner, date.
• Indexed asynchronously — there's a lag between data change and index update (typically seconds to minutes).
• Specific configured fields — admins choose which tables and which fields are indexed.

Quick Find.

The traditional search box on each view. Built directly on Dataverse's SQL search.

Characteristics:

• Always current — searches the live database; no indexing lag.
• Per-table only — searches the current view's underlying table.
• Exact-match-oriented — wildcard matching (query) on specific columns; less forgiving than Dataverse search.
• Slower at scale — direct DB scan on large tables can be slow.
• No relevance ranking — results returned in the view's sort order.
• Configurable per table — Quick Find View per table defines which columns to search.

When to use which.

• Dataverse search for: global discovery ("find anything about Karen Smith"), cross-table search, typo tolerance, ranked results, modern UX.
• Quick Find for: live-current data (no indexing lag), per-view filtering, specific column-bounded search, simple exact-match needs.

Configuration — Dataverse search.

Per environment:

• Enabled or disabled at the environment level.
• Tables enabled for search — admins pick which tables are indexed. Default includes Accounts, Contacts, Activities, etc.; custom tables need explicit enabling.
• Columns enabled — per table, which columns are indexed. Names, descriptions, primary fields by default; admins extend.
• Privileges — security roles control which users have access to Dataverse search.

After enabling new tables or columns, an initial index has to build — can take hours for large tables.

Configuration — Quick Find.

Per table:

• Quick Find View — a special view that defines:
  • Which columns are searched.
  • Which columns are returned in results.
  • Default sort order.

Editing the Quick Find View adds or removes searchable columns.

Common pitfalls.

• Dataverse search not enabled — users complain that "search doesn't find anything"; admins haven't enabled it on relevant tables.
• Wrong columns indexed — searches don't return what users expect because the field they're searching isn't indexed.
• Stale index — recent record changes don't appear in Dataverse search results because the index hasn't refreshed. Usually transient; chronic lag suggests an indexing issue.
• Quick Find too broad — searching too many columns slows the query. Tune to the columns users actually search.

The mobile experience

The Power Apps mobile app uses Dataverse search by default for its global search box. The seller in the field saying "search doesn't work" is almost always a Dataverse search configuration issue, not a mobile-specific problem.

Power Apps canvas apps

Canvas apps can call both search mechanisms through Dataverse connectors. For typed-ahead search-as-you-type, Dataverse search is the right pattern; for narrow filtered list searches, the standard Filter() function on a Dataverse table is sufficient.

Operational discipline

Enable Dataverse search on every meaningful table; verify the right columns are indexed; ensure users have appropriate security to use it. Quick Find tunes per-table for users' common searches. Both are configuration, not customisation — get them right early.