Find the actual unworked count
Group records by order and distinguish records never reviewed from records with documented attempts or agreed next steps. An unanswered attempt is work performed, but it is not a conversation. Check whether notes live in an approved external workspace before labeling a record untouched. Use one consistent definition across the team.
Look for the operational cause
Review recent buying volume against available agent time. Check failed imports, unclear ownership, absences and oversized batches. If an order was delivered but never assigned, improving a script will not solve the immediate problem. Identify the point where records stopped moving and assign an owner to repair that process.
Protect existing commitments
Reserve time for requested callbacks and appointments before allocating recovery work. Review each record’s context and current contact instructions before any new attempt. Do not compensate for backlog by applying an arbitrary high-volume outreach schedule. The appropriate next action depends on the record and your approved contact process.
Reduce intake until the queue is manageable
Set a daily review capacity based on actual handling time and track the remaining unworked count. Record why records are closed or removed rather than making the backlog disappear without explanation. Resume larger purchases only when the team can handle existing commitments and new work. A smaller, understood queue is easier to learn from than an expanding pile of unexplained records.