Ask a question and Claude selects from four read-only tools. You do not need to call them by name.
Whose recordings are visible, and what the filters accept.
Reports whose recordings this connection can read, how many there are, the date range they cover, and the exact call-type and rep names that the other tools accept as filters. Claude calls this first, because guessing a label is the most common reason a query comes back empty when data exists.
Recordings newest first, with filters. Metadata only.
Lists recordings newest first, filterable by call type, rep and date range. It returns metadata without transcript text. A filter that matches nothing in the workspace is reported back rather than silently returning an empty list, so a misspelled name never looks like an absence of calls.
What was said on one recording, as speaker-attributed turns.
Returns what was said on a single recording as speaker-attributed turns, using TENEKS’s own speaker attribution where it has identified who was talking. Long calls are paged, so Claude can read a two-hour call without truncating it.
A grounded answer to a question across many recordings.
Answers questions about recurring objections, rep activity or changes across several recordings, using the calls as sources. It can come back marked incomplete, which means the answer could not be fully grounded and should not be repeated as fact.